DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Claim Objections
Claims 2-3 and 12-13 are objected to because of the following informalities:
In claim 2, lines 3-4, “the evaluation results” should read “the evaluation result” to properly reference “an evaluation result” in line 7 of claim 1
In claim 12, lines 3-4, “the evaluation results” should read “the evaluation result” to properly reference “an evaluation result” in line 10 of claim 11
Dependent claim 3 is objected based on being directly or indirectly dependent on objected claim 2.
Dependent claim 13 is objected based on being directly or indirectly dependent on objected claim 12.
Appropriate correction is required.
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Regarding Claim 1,
Claim 1 is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 1 is directed to a method, which is directed to a process, one of the statutory categories.
Step 2A Prong One Analysis: The limitations:
“extracting, …, logic rules from the trained link prediction model”
“generating, …, at least one evaluation metric based on the logic rules”
“generating, …, an evaluation result based on comparing the at least one evaluation metric to a predetermined criteria”
As drafted, under their broadest reasonable interpretations, cover mental processes (concepts performed in the human mind (including an observation, evaluation, judgement, opinion)) but for the recitation of mere instructions to apply language (See MPEP 2106.05(f)) and insignificant extra-solution activity (See MPEP 2106.05(g)). The above limitations in the context of this claim encompass extracting logic rules from the trained link prediction model (corresponds to evaluation and judgement; in particular, a human, with the assistance of pen and paper, can extract logic rules from the trained link prediction model); generating at least one evaluation metric based on the logic rules (corresponds to evaluation and judgement; in particular, a human, with the assistance of pen and paper, can use the logic rules to generate an evaluation metric); and generating an evaluation result based on comparing the evaluation metric to a predetermined criteria (corresponds to evaluation and judgement; in particular, a human, with the assistance of pen and paper, can compare the evaluation metric to a predetermined criteria to generate an evaluation result).
Step 2A Prong Two Analysis: The judicial exceptions are not integrated into a practical application. In particular, the claim recites additional elements that are mere instructions to apply (See MPEP 2106.05(f)) or insignificant extra-solution activity (See MPEP 2106.05(g)). The limitations:
“an artificial intelligence (Al) device”
“a processor”
“training, via the processor, a link prediction model on the knowledge graph to generate a trained link prediction model”
“via the processor”
As drafted, are additional elements that amount to no more than mere instructions to apply the exception for the abstract ideas. See MPEP 2106.05(f). The limitations:
“obtaining, via a processor in the Al device, a knowledge graph”
“outputting, via an output unit in the Al device, the evaluation result”
As drafted, are additional elements that correspond to insignificant extra-solution activity. In particular, the additional elements are merely directed towards mere data gathering. See MPEP 2106.05(g). Therefore, the additional elements do not integrate the abstract ideas into a practical application.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to integration of the abstract idea into a practical application, all of the additional elements are “mere instructions to apply an exception” (I.e. the additional elements describe a generic AI device, processor, and generic model training for applying the abstract ideas) or insignificant extra-solution activity (i.e. obtaining/receiving and outputting/transmitting data). Furthermore, the “obtaining …” and “outputting …” limitations are insignificant extra-solution activity that is well-understood, routine, and conventional according to MPEP 2106.05(d) (“The courts have recognized the following computer functions as well‐understood, routine, and conventional functions when they are claimed in a merely generic manner (e.g., at a high level of generality) or as insignificant extra-solution activity… i. Receiving or transmitting data over a network). Mere instructions to apply an exception cannot provide an inventive concept. The claim is not patent eligible.
Regarding Claim 2,
Claim 2 is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 2 is directed to a method, which is directed to a process, one of the statutory categories.
Step 2A Prong One Analysis: Please see the analysis of claim 1. The limitations of claim 2 are only additional elements to the abstract ideas of claim 1.
Step 2A Prong Two Analysis: The judicial exceptions are not integrated into a practical application. In particular, the claim recites additional elements that are mere instructions to apply (See MPEP 2106.05(f)) or insignificant extra-solution activity (See MPEP 2106.05(g)). The limitations:
“saving the trained link prediction model in a memory of the Al device for deployment or transmitting the trained link prediction model to an external device for deployment, based on the evaluation results”
As drafted, are additional elements that correspond to insignificant extra-solution activity. In particular, the additional elements are merely directed towards mere data gathering. See MPEP 2106.05(g). In addition, the recitation of additional elements in claim 1 of a generic AI device, processor, and generic model training, as drafted, are reciting mere instructions to apply language such that it amounts to no more than mere instructions to apply the exceptions. Furthermore, the “obtaining …” and “outputting …” limitations of claim 1 are additional elements that correspond to insignificant extra-solution activity as mere data gathering. Therefore, the additional elements do not integrate the abstract ideas into a practical application.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to integration of the abstract idea into a practical application, all of the additional elements are “mere instructions to apply an exception” (I.e. the additional elements describe a generic AI device, processor, and generic model training for applying the abstract ideas) or insignificant extra-solution activity (i.e. obtaining/receiving, outputting/transmitting, and saving/storing data). Furthermore, the “obtaining …”, “outputting …”, “saving …”, and “transmitting …” limitations are insignificant extra-solution activity that is well-understood, routine, and conventional according to MPEP 2106.05(d) (“The courts have recognized the following computer functions as well‐understood, routine, and conventional functions when they are claimed in a merely generic manner (e.g., at a high level of generality) or as insignificant extra-solution activity… i. Receiving or transmitting data over a network … iv. Storing and retrieving information in memory). Mere instructions to apply an exception cannot provide an inventive concept. The claim is not patent eligible.
Regarding Claim 3,
Claim 3 is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 3 is directed to a method, which is directed to a process, one of the statutory categories.
Step 2A Prong One Analysis: Please see the analysis of claim 2. The limitations of claim 3 are only additional elements to the abstract ideas of claim 2.
Step 2A Prong Two Analysis: The judicial exceptions are not integrated into a practical application. In particular, the claim recites additional elements that are mere instructions to apply (See MPEP 2106.05(f)), generally linking the use of a judicial exception to a particular technological environment or field of use (See MPEP 2106.05(h)), or insignificant extra-solution activity (See MPEP 2106.05(g)). The limitation:
“wherein the trained link prediction model is deployed in a question and answer system or a recommendation system”
As drafted, is an additional element that amounts to generally linking the use of a judicial exception to a particular technological environment or field of use (QA/recommendation system). See MPEP 2106.05(h). In addition, the recitation of additional elements in claim 2 of a generic AI device, processor, and generic model training, as drafted, are reciting mere instructions to apply language such that it amounts to no more than mere instructions to apply the exceptions. Furthermore, the “obtaining …”, “outputting …”, “saving …”, and “transmitting …” limitations of claim 2 are additional elements that correspond to insignificant extra-solution activity as mere data gathering. Therefore, the additional elements do not integrate the abstract ideas into a practical application.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to integration of the abstract idea into a practical application, all of the additional elements are generally linking the use of a judicial exception to a particular technological environment or field of use (QA/recommendation system), or are “mere instructions to apply an exception” (I.e. the additional elements describe a generic AI device, processor, and generic model training for applying the abstract ideas) or insignificant extra-solution activity (i.e. obtaining/receiving, outputting/transmitting, and saving/storing data). Furthermore, the “obtaining …”, “outputting …”, “saving …”, and “transmitting …” limitations are insignificant extra-solution activity that is well-understood, routine, and conventional according to MPEP 2106.05(d) (“The courts have recognized the following computer functions as well‐understood, routine, and conventional functions when they are claimed in a merely generic manner (e.g., at a high level of generality) or as insignificant extra-solution activity… i. Receiving or transmitting data over a network … iv. Storing and retrieving information in memory). Mere instructions to apply an exception cannot provide an inventive concept. The claim is not patent eligible.
Regarding Claim 4,
Claim 4 is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 4 is directed to a method, which is directed to a process, one of the statutory categories.
Step 2A Prong One Analysis: The limitations:
“generating a rules recovered metric, the rules recovered metric being a value based on dividing a number of the logic rules by a total number of original logic rules for the knowledge graph”
“generating a graph coverage metric, the graph coverage metric being a value based on dividing a number of grounded paths covered by the logic rules by a total number of ground paths for the knowledge graph”
As drafted, under their broadest reasonable interpretations, cover mental processes (concepts performed in the human mind (including an observation, evaluation, judgement, opinion)) but for the recitation of mere instructions to apply language (See MPEP 2106.05(f)) and insignificant extra-solution activity (See MPEP 2106.05(g)). The above limitations in the context of this claim encompass generating a rules recovered metric that is a value based on dividing a number of the logic rules by a total number of original logic rules for the knowledge graph (corresponds to evaluation and judgement; in particular, a human, with the assistance of pen and paper, can generate a rules recovered metric by dividing a number of the logic rules by a total number of original logic rules for the knowledge graph); and generating a graph coverage metric that is a value based on dividing a number of grounded paths covered by the logic rules by a total number of ground paths for the knowledge graph (corresponds to evaluation and judgement; in particular, a human, with the assistance of pen and paper, can generate a graph coverage metric by dividing a number of grounded paths covered by the logic rules by a total number of ground paths for the knowledge graph).
Step 2A Prong Two Analysis: The judicial exceptions are not integrated into a practical application. In particular, the claim recites additional elements that are mere instructions to apply (See MPEP 2106.05(f)) or insignificant extra-solution activity (See MPEP 2106.05(g)). The recitation of additional elements in claim 1 of a generic AI device, processor, and generic model training, as drafted, are reciting mere instructions to apply language such that it amounts to no more than mere instructions to apply the exceptions. Furthermore, the “obtaining …” and “outputting …” limitations of claim 1 are additional elements that correspond to insignificant extra-solution activity as mere data gathering. Therefore, the additional elements do not integrate the abstract ideas into a practical application.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to integration of the abstract idea into a practical application, all of the additional elements are “mere instructions to apply an exception” (I.e. the additional elements describe a generic AI device, processor, and generic model training for applying the abstract ideas) or insignificant extra-solution activity (i.e. obtaining/receiving and outputting/transmitting data). Furthermore, the “obtaining …” and “outputting …” limitations are insignificant extra-solution activity that is well-understood, routine, and conventional according to MPEP 2106.05(d) (“The courts have recognized the following computer functions as well‐understood, routine, and conventional functions when they are claimed in a merely generic manner (e.g., at a high level of generality) or as insignificant extra-solution activity… i. Receiving or transmitting data over a network). Mere instructions to apply an exception cannot provide an inventive concept. The claim is not patent eligible.
Regarding Claim 5,
Claim 5 is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 5 is directed to a method, which is directed to a process, one of the statutory categories.
Step 2A Prong One Analysis: The limitations:
“comparing the rules recovered metric to a first predetermined threshold”
“comparing the graph coverage metric to a second predetermined threshold”
As drafted, under their broadest reasonable interpretations, cover mental processes (concepts performed in the human mind (including an observation, evaluation, judgement, opinion)) and mathematical concepts (mathematical relationships, mathematical formulas or equations, mathematical calculations) but for the recitation of mere instructions to apply language (See MPEP 2106.05(f)) and insignificant extra-solution activity (See MPEP 2106.05(g)). The above limitations in the context of this claim encompass comparing the rules covered metric to a first predetermined threshold (corresponds to evaluation and judgement; in particular, a human, with the assistance of pen and paper, can compare the rules recovered metric to a first predetermined threshold); and comparing the graph coverage metric to a second predetermined threshold (corresponds to evaluation and judgement; in particular, a human, with the assistance of pen and paper, can compare the graph coverage metric to a second predetermined threshold).
Step 2A Prong Two Analysis: The judicial exceptions are not integrated into a practical application. In particular, the claim recites additional elements that are mere instructions to apply (See MPEP 2106.05(f)) or insignificant extra-solution activity (See MPEP 2106.05(g)). The limitations:
“in response to the rules recovered metric being greater than or equal to the first predetermined threshold and the graph coverage metric being greater than or equal to the second predetermined threshold, saving the trained link prediction model in a memory of the Al device for deployment or transmitting the trained link prediction model to an external device for deployment”
As drafted, are additional elements that correspond to insignificant extra-solution activity. In particular, the additional elements are merely directed towards mere data gathering. See MPEP 2106.05(g). In addition, the recitation of additional elements in claim 4 of a generic AI device, processor, and generic model training, as drafted, are reciting mere instructions to apply language such that it amounts to no more than mere instructions to apply the exceptions. Furthermore, the “obtaining …” and “outputting …” limitations of claim 4 are additional elements that correspond to insignificant extra-solution activity as mere data gathering. Therefore, the additional elements do not integrate the abstract ideas into a practical application.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to integration of the abstract idea into a practical application, all of the additional elements are “mere instructions to apply an exception” (I.e. the additional elements describe a generic AI device, processor, and generic model training for applying the abstract ideas) or insignificant extra-solution activity (i.e. obtaining/receiving, outputting/transmitting, and saving/storing data). Furthermore, the “obtaining …”, “outputting …”, “… saving …”, and “… transmitting …” limitations are insignificant extra-solution activity that is well-understood, routine, and conventional according to MPEP 2106.05(d) (“The courts have recognized the following computer functions as well‐understood, routine, and conventional functions when they are claimed in a merely generic manner (e.g., at a high level of generality) or as insignificant extra-solution activity… i. Receiving or transmitting data over a network … iv. Storing and retrieving information in memory). Mere instructions to apply an exception cannot provide an inventive concept. The claim is not patent eligible.
Regarding Claim 6,
Claim 6 is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 6 is directed to a method, which is directed to a process, one of the statutory categories.
Step 2A Prong One Analysis: The limitation:
“generating one quality score for the logic rules based on a sum of the plurality of scores and based on dividing by a total number of the logic rules”
As drafted, under their broadest reasonable interpretations, cover mental processes (concepts performed in the human mind (including an observation, evaluation, judgement, opinion)) but for the recitation of mere instructions to apply language (See MPEP 2106.05(f)) and insignificant extra-solution activity (See MPEP 2106.05(g)). The above limitations in the context of this claim encompass generating a quality score for the logic rules based on a sum of the plurality of scores based on dividing by a total number of the logic rules (corresponds to evaluation and judgement; in particular, a human, with the assistance of pen and paper, can generate one quality score for the logic rules based on a sum of the plurality of quality scores and based on dividing by a total number of the logic rules).
Step 2A Prong Two Analysis: The judicial exceptions are not integrated into a practical application. In particular, the claim recites additional elements that are mere instructions to apply (See MPEP 2106.05(f)) or insignificant extra-solution activity (See MPEP 2106.05(g)). The limitation:
“receiving a plurality of scores for each logic rule among the logic rules, the plurality of scores being assigned by a plurality of annotators”
As drafted, is an additional element that corresponds to insignificant extra-solution activity. In particular, the additional elements are merely directed towards mere data gathering. See MPEP 2106.05(g). In addition, the recitation of additional elements in claim 1 of a generic AI device, processor, and generic model training, as drafted, are reciting mere instructions to apply language such that it amounts to no more than mere instructions to apply the exceptions. Furthermore, the “obtaining …” and “outputting …” limitations of claim 1 are additional elements that correspond to insignificant extra-solution activity as mere data gathering. Therefore, the additional elements do not integrate the abstract ideas into a practical application.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to integration of the abstract idea into a practical application, all of the additional elements are “mere instructions to apply an exception” (I.e. the additional elements describe a generic AI device, processor, and generic model training for applying the abstract ideas) or insignificant extra-solution activity (i.e. obtaining/receiving and outputting/transmitting data). Furthermore, the “obtaining …”, “outputting …”, and “receiving …” limitations are insignificant extra-solution activity that is well-understood, routine, and conventional according to MPEP 2106.05(d) (“The courts have recognized the following computer functions as well‐understood, routine, and conventional functions when they are claimed in a merely generic manner (e.g., at a high level of generality) or as insignificant extra-solution activity… i. Receiving or transmitting data over a network). Mere instructions to apply an exception cannot provide an inventive concept. The claim is not patent eligible.
Regarding Claim 7,
Claim 7 is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 7 is directed to a method, which is directed to a process, one of the statutory categories.
Step 2A Prong One Analysis: Please see the analysis of claim 6. The limitations of claim 7 are only additional elements to the abstract ideas of claim 6.
Step 2A Prong Two Analysis: The judicial exceptions are not integrated into a practical application. In particular, the claim recites additional elements that are mere instructions to apply (See MPEP 2106.05(f)) or insignificant extra-solution activity (See MPEP 2106.05(g)). The limitation:
“wherein each of the plurality of scores is a binary value of 1 or 0”
As drafted, is part of the insignificant extra-solution activity of claim 6. The limitation of claim 7 further limits the limitation of claim 6 by further defining what the received plurality of scores comprises. In addition, the recitation of additional elements in claim 6 of a generic AI device, processor, and generic model training, as drafted, are reciting mere instructions to apply language such that it amounts to no more than mere instructions to apply the exceptions. Furthermore, the “obtaining …”, “outputting …”, and “receiving …” limitations of claim 6 are additional elements that correspond to insignificant extra-solution activity as mere data gathering. Therefore, the additional elements do not integrate the abstract ideas into a practical application.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to integration of the abstract idea into a practical application, all of the additional elements are “mere instructions to apply an exception” (I.e. the additional elements describe a generic AI device, processor, and generic model training for applying the abstract ideas) or insignificant extra-solution activity (i.e. obtaining/receiving and outputting/transmitting data). Furthermore, the “obtaining …”, “outputting …”, and “receiving …” limitations are insignificant extra-solution activity that is well-understood, routine, and conventional according to MPEP 2106.05(d) (“The courts have recognized the following computer functions as well‐understood, routine, and conventional functions when they are claimed in a merely generic manner (e.g., at a high level of generality) or as insignificant extra-solution activity… i. Receiving or transmitting data over a network). Mere instructions to apply an exception cannot provide an inventive concept. The claim is not patent eligible.
Regarding Claim 8,
Claim 8 is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 8 is directed to a method, which is directed to a process, one of the statutory categories.
Step 2A Prong One Analysis: The limitation:
“comparing the one quality score to a third predetermined threshold”
As drafted, under their broadest reasonable interpretations, cover mental processes (concepts performed in the human mind (including an observation, evaluation, judgement, opinion)) but for the recitation of mere instructions to apply language (See MPEP 2106.05(f)) and insignificant extra-solution activity (See MPEP 2106.05(g)). The above limitations in the context of this claim encompass comparing the one quality score to a third predetermined threshold (corresponds to evaluation and judgement; in particular, a human, with the assistance of pen and paper, can compare the one quality score to a third predetermined threshold).
Step 2A Prong Two Analysis: The judicial exceptions are not integrated into a practical application. In particular, the claim recites additional elements that are mere instructions to apply (See MPEP 2106.05(f)) or insignificant extra-solution activity (See MPEP 2106.05(g)). The limitations:
“in response to the one quality score being greater than or equal to the third predetermined threshold, saving the trained link prediction model in a memory of the Al device for deployment or transmitting the trained link prediction model to an external device for deployment”
As drafted, are additional elements that correspond to insignificant extra-solution activity. In particular, the additional elements are merely directed towards mere data gathering. See MPEP 2106.05(g). In addition, the recitation of additional elements in claim 6 of a generic AI device, processor, and generic model training, as drafted, are reciting mere instructions to apply language such that it amounts to no more than mere instructions to apply the exceptions. Furthermore, the “obtaining …”, “outputting …”, and “receiving …” limitations of claim 6 are additional elements that correspond to insignificant extra-solution activity as mere data gathering. Therefore, the additional elements do not integrate the abstract ideas into a practical application.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to integration of the abstract idea into a practical application, all of the additional elements are “mere instructions to apply an exception” (I.e. the additional elements describe a generic AI device, processor, and generic model training for applying the abstract ideas) or insignificant extra-solution activity (i.e. obtaining/receiving, outputting/transmitting, and saving/storing data). Furthermore, the “obtaining …”, “outputting …”, “receiving …”, “… saving …”, and “… transmitting …” limitations are insignificant extra-solution activity that is well-understood, routine, and conventional according to MPEP 2106.05(d) (“The courts have recognized the following computer functions as well‐understood, routine, and conventional functions when they are claimed in a merely generic manner (e.g., at a high level of generality) or as insignificant extra-solution activity… i. Receiving or transmitting data over a network … iv. Storing and retrieving information in memory). Mere instructions to apply an exception cannot provide an inventive concept. The claim is not patent eligible.
Regarding Claim 9,
Claim 9 is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 9 is directed to a method, which is directed to a process, one of the statutory categories.
Step 2A Prong One Analysis: The limitations:
“generating link prediction metrics based on the trained link prediction model, the link prediction metrics including at least one of mean rank (MR), mean reciprocal rank (MRR), and Hit@K”
“determining whether to deploy the trained link prediction model based on the link prediction metrics”
As drafted, under their broadest reasonable interpretations, cover mental processes (concepts performed in the human mind (including an observation, evaluation, judgement, opinion)) but for the recitation of mere instructions to apply language (See MPEP 2106.05(f)) and insignificant extra-solution activity (See MPEP 2106.05(g)). The above limitations in the context of this claim encompass generating link prediction metric based on the trained link prediction model, the link prediction metrics including at least one of MR, MMR, and Hit@K (corresponds to evaluation and judgement; in particular, a human, with the assistance of pen and paper, can generate link prediction metric including at least one of MR, MMR, and Hit@K based on the trained link prediction model); and determining whether to deploy the trained link prediction model based on the link prediction metrics (corresponds to evaluation and judgement; in particular, a human, with the assistance of pen and paper, can use the link prediction metrics to determine whether to deploy the trained link prediction model).
Step 2A Prong Two Analysis: The judicial exceptions are not integrated into a practical application. In particular, the claim recites additional elements that are mere instructions to apply (See MPEP 2106.05(f)) or insignificant extra-solution activity (See MPEP 2106.05(g)). The recitation of additional elements in claim 1 of a generic AI device, processor, and generic model training, as drafted, are reciting mere instructions to apply language such that it amounts to no more than mere instructions to apply the exceptions. Furthermore, the “obtaining …” and “outputting …” limitations of claim 1 are additional elements that correspond to insignificant extra-solution activity as mere data gathering. Therefore, the additional elements do not integrate the abstract ideas into a practical application.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to integration of the abstract idea into a practical application, all of the additional elements are “mere instructions to apply an exception” (I.e. the additional elements describe a generic AI device, processor, and generic model training for applying the abstract ideas) or insignificant extra-solution activity (i.e. obtaining/receiving and outputting/transmitting data). Furthermore, the “obtaining …” and “outputting …” limitations are insignificant extra-solution activity that is well-understood, routine, and conventional according to MPEP 2106.05(d) (“The courts have recognized the following computer functions as well‐understood, routine, and conventional functions when they are claimed in a merely generic manner (e.g., at a high level of generality) or as insignificant extra-solution activity… i. Receiving or transmitting data over a network). Mere instructions to apply an exception cannot provide an inventive concept. The claim is not patent eligible.
Regarding Claim 10,
Claim 10 is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 10 is directed to a method, which is directed to a process, one of the statutory categories.
Step 2A Prong One Analysis: Please see the analysis of claim 1. The limitations of claim 10 are only additional elements to the abstract ideas of claim 1.
Step 2A Prong Two Analysis: The judicial exceptions are not integrated into a practical application. In particular, the claim recites additional elements that are mere instructions to apply (See MPEP 2106.05(f)) or insignificant extra-solution activity (See MPEP 2106.05(g)). The limitation:
“wherein the Al device includes at least one of a smart television, a mobile phone, and a home appliance device”
As drafted, is an additional element that amounts to no more than mere instructions to apply the exception for the abstract ideas. See MPEP 2106.05(f). In addition, the recitation of additional elements in claim 1 of a generic AI device, processor, and generic model training, as drafted, are reciting mere instructions to apply language such that it amounts to no more than mere instructions to apply the exceptions. Furthermore, the “obtaining …” and “outputting …” limitations of claim 1 are additional elements that correspond to insignificant extra-solution activity as mere data gathering. Therefore, the additional elements do not integrate the abstract ideas into a practical application.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to integration of the abstract idea into a practical application, all of the additional elements are “mere instructions to apply an exception” (I.e. the additional elements describe a generic AI device, processor, and generic model training for applying the abstract ideas) or insignificant extra-solution activity (i.e. obtaining/receiving and outputting/transmitting data). Furthermore, the “obtaining …” and “outputting …” limitations are insignificant extra-solution activity that is well-understood, routine, and conventional according to MPEP 2106.05(d) (“The courts have recognized the following computer functions as well‐understood, routine, and conventional functions when they are claimed in a merely generic manner (e.g., at a high level of generality) or as insignificant extra-solution activity… i. Receiving or transmitting data over a network). Mere instructions to apply an exception cannot provide an inventive concept. The claim is not patent eligible.
Regarding Claim 11,
Claim 11 is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 11 is directed to a device, which is directed to a machine, one of the statutory categories.
Step 2A Prong One Analysis: The limitations:
“extract logic rules from the trained link prediction model”
“generate at least one evaluation metric based on the logic rules”
“generate an evaluation result based on comparing the at least one evaluation metric to a predetermined criteria”
As drafted, under their broadest reasonable interpretations, cover mental processes (concepts performed in the human mind (including an observation, evaluation, judgement, opinion)) but for the recitation of mere instructions to apply language (See MPEP 2106.05(f)) and insignificant extra-solution activity (See MPEP 2106.05(g)). The above limitations in the context of this claim encompass extracting logic rules from the trained link prediction model (corresponds to evaluation and judgement; in particular, a human, with the assistance of pen and paper, can extract logic rules from the trained link prediction model); generating at least one evaluation metric based on the logic rules (corresponds to evaluation and judgement; in particular, a human, with the assistance of pen and paper, can use the logic rules to generate an evaluation metric); and generating an evaluation result based on comparing the evaluation metric to a predetermined criteria (corresponds to evaluation and judgement; in particular, a human, with the assistance of pen and paper, can compare the evaluation metric to a predetermined criteria to generate an evaluation result).
Step 2A Prong Two Analysis: The judicial exceptions are not integrated into a practical application. In particular, the claim recites additional elements that are mere instructions to apply (See MPEP 2106.05(f)) or insignificant extra-solution activity (See MPEP 2106.05(g)). The limitations:
“a memory”
“a controller”
“train a link prediction model on the knowledge graph to generate a trained link prediction model”
As drafted, are additional elements that amount to no more than mere instructions to apply the exception for the abstract ideas. See MPEP 2106.05(f). The limitations:
“obtain a knowledge graph”
“output the evaluation result”
As drafted, are additional elements that correspond to insignificant extra-solution activity. In particular, the additional elements are merely directed towards mere data gathering. See MPEP 2106.05(g). Therefore, the additional elements do not integrate the abstract ideas into a practical application.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to integration of the abstract idea into a practical application, all of the additional elements are “mere instructions to apply an exception” (I.e. the additional elements describe a generic memory, controller, and generic model training for applying the abstract ideas) or insignificant extra-solution activity (i.e. obtaining/receiving and outputting/transmitting data). Furthermore, the “obtain …” and “output …” limitations are insignificant extra-solution activity that is well-understood, routine, and conventional according to MPEP 2106.05(d) (“The courts have recognized the following computer functions as well‐understood, routine, and conventional functions when they are claimed in a merely generic manner (e.g., at a high level of generality) or as insignificant extra-solution activity… i. Receiving or transmitting data over a network). Mere instructions to apply an exception cannot provide an inventive concept. The claim is not patent eligible.
Regarding Claim 12,
Claim 12 is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 12 is directed to a device, which is directed to a machine, one of the statutory categories.
Step 2A Prong One Analysis: Please see the analysis of claim 11. The limitations of claim 12 are only additional elements to the abstract ideas of claim 11.
Step 2A Prong Two Analysis: The judicial exceptions are not integrated into a practical application. In particular, the claim recites additional elements that are mere instructions to apply (See MPEP 2106.05(f)) or insignificant extra-solution activity (See MPEP 2106.05(g)). The limitation:
“the controller”
As drafted, is an additional element that amounts to no more than mere instructions to apply the exception for the abstract ideas. See MPEP 2106.05(f). The limitations:
“save the trained link prediction model in the memory of the Al device for deployment or transmit the trained link prediction model to an external device for deployment, based on the evaluation results”
As drafted, are additional elements that correspond to insignificant extra-solution activity. In particular, the additional elements are merely directed towards mere data gathering. See MPEP 2106.05(g). In addition, the recitation of additional elements in claim 11 of a generic memory, controller, and generic model training, as drafted, are reciting mere instructions to apply language such that it amounts to no more than mere instructions to apply the exceptions. Furthermore, the “obtaining …” and “outputting …” limitations of claim 11 are additional elements that correspond to insignificant extra-solution activity as mere data gathering. Therefore, the additional elements do not integrate the abstract ideas into a practical application.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to integration of the abstract idea into a practical application, all of the additional elements are “mere instructions to apply an exception” (I.e. the additional elements describe a generic memory, controller, and generic model training for applying the abstract ideas) or insignificant extra-solution activity (i.e. obtaining/receiving, outputting/transmitting, and saving/storing data). Furthermore, the “obtain …”, “output …”, “save …”, and “transmit …” limitations are insignificant extra-solution activity that is well-understood, routine, and conventional according to MPEP 2106.05(d) (“The courts have recognized the following computer functions as well‐understood, routine, and conventional functions when they are claimed in a merely generic manner (e.g., at a high level of generality) or as insignificant extra-solution activity… i. Receiving or transmitting data over a network … iv. Storing and retrieving information in memory). Mere instructions to apply an exception cannot provide an inventive concept. The claim is not patent eligible.
Regarding Claim 13,
Claim 13 is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 13 is directed to a device, which is directed to a machine, one of the statutory categories.
Step 2A Prong One Analysis: Please see the analysis of claim 12. The limitations of claim 13 are only additional elements to the abstract ideas of claim 12.
Step 2A Prong Two Analysis: The judicial exceptions are not integrated into a practical application. In particular, the claim recites additional elements that are mere instructions to apply (See MPEP 2106.05(f)), generally linking the use of a judicial exception to a particular technological environment or field of use (See MPEP 2106.05(h)), or insignificant extra-solution activity (See MPEP 2106.05(g)). The limitation:
“wherein the trained link prediction model is deployed in a question and answer system or a recommendation system”
As drafted, is an additional element that amounts to generally linking the use of a judicial exception to a particular technological environment or field of use (QA/recommendation system). See MPEP 2106.05(h). In addition, the recitation of additional elements in claim 12 of a generic memory, controller, and generic model training, as drafted, are reciting mere instructions to apply language such that it amounts to no more than mere instructions to apply the exceptions. Furthermore, the “obtain …”, “output …”, “save …”, and “transmit …” limitations of claim 12 are additional elements that correspond to insignificant extra-solution activity as mere data gathering. Therefore, the additional elements do not integrate the abstract ideas into a practical application.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to integration of the abstract idea into a practical application, all of the additional elements are generally linking the use of a judicial exception to a particular technological environment or field of use (QA/recommendation system), or are “mere instructions to apply an exception” (I.e. the additional elements describe a generic memory, controller, and generic model training for applying the abstract ideas) or insignificant extra-solution activity (i.e. obtaining/receiving, outputting/transmitting, and saving/storing data). Furthermore, the “obtain …”, “output …”, “save …”, and “transmit …” limitations are insignificant extra-solution activity that is well-understood, routine, and conventional according to MPEP 2106.05(d) (“The courts have recognized the following computer functions as well‐understood, routine, and conventional functions when they are claimed in a merely generic manner (e.g., at a high level of generality) or as insignificant extra-solution activity… i. Receiving or transmitting data over a network … iv. Storing and retrieving information in memory). Mere instructions to apply an exception cannot provide an inventive concept. The claim is not patent eligible.
Regarding Claim 14,
Claim 14 is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 14 is directed to a device, which is directed to a machine, one of the statutory categories.
Step 2A Prong One Analysis: The limitations:
“generate a rules recovered metric, the rules recovered metric being a value based on dividing a number of the logic rules by a total number of original logic rules for the knowledge graph”
“generate a graph coverage metric, the graph coverage metric being a value based on dividing a number of grounded paths covered by the logic rules by a total number of ground paths for the knowledge graph”
As drafted, generating a rules recovered metric that is a value based on dividing a number of the logic rules by a total number of original logic rules for the knowledge graph (corresponds to evaluation and judgement; in particular, a human, with the assistance of pen and paper, can generate a rules recovered metric by dividing a number of the logic rules by a total number of original logic rules for the knowledge graph); and generating a graph coverage metric that is a value based on dividing a number of grounded paths covered by the logic rules by a total number of ground paths for the knowledge graph (corresponds to evaluation and judgement; in particular, a human, with the assistance of pen and paper, can generate a graph coverage metric by dividing a number of grounded paths covered by the logic rules by a total number of ground paths for the knowledge graph).
Step 2A Prong Two Analysis: The judicial exceptions are not integrated into a practical application. In particular, the claim recites additional elements that are mere instructions to apply (See MPEP 2106.05(f)) or insignificant extra-solution activity (See MPEP 2106.05(g)). The limitation:
“the controller”
As drafted, is an additional element that amounts to no more than mere instructions to apply the exception for the abstract ideas. See MPEP 2106.05(f). In addition, the recitation of additional elements in claim 11 of a generic memory, controller, and generic model training, as drafted, are reciting mere instructions to apply language such that it amounts to no more than mere instructions to apply the exceptions. Furthermore, the “obtaining …” and “outputting …” limitations of claim 11 are additional elements that correspond to insignificant extra-solution activity as mere data gathering. Therefore, the additional elements do not integrate the abstract ideas into a practical application.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to integration of the abstract idea into a practical application, all of the additional elements are “mere instructions to apply an exception” (I.e. the additional elements describe a generic memory, controller, and generic model training for applying the abstract ideas) or insignificant extra-solution activity (i.e. obtaining/receiving and outputting/transmitting data). Furthermore, the “obtain …” and “output …” limitations are insignificant extra-solution activity that is well-understood, routine, and conventional according to MPEP 2106.05(d) (“The courts have recognized the following computer functions as well‐understood, routine, and conventional functions when they are claimed in a merely generic manner (e.g., at a high level of generality) or as insignificant extra-solution activity… i. Receiving or transmitting data over a network). Mere instructions to apply an exception cannot provide an inventive concept. The claim is not patent eligible.
Regarding Claim 15,
Claim 15 is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 15 is directed to a device, which is directed to a machine, one of the statutory categories.
Step 2A Prong One Analysis: The limitations:
“compare the rules recovered metric to a first predetermined threshold”
“compare the graph coverage metric to a second predetermined threshold”
As drafted, under their broadest reasonable interpretations, cover mental processes (concepts performed in the human mind (including an observation, evaluation, judgement, opinion)) and mathematical concepts (mathematical relationships, mathematical formulas or equations, mathematical calculations) but for the recitation of mere instructions to apply language (See MPEP 2106.05(f)) and insignificant extra-solution activity (See MPEP 2106.05(g)). The above limitations in the context of this claim encompass comparing the rules covered metric to a first predetermined threshold (corresponds to evaluation and judgement; in particular, a human, with the assistance of pen and paper, can compare the rules recovered metric to a first predetermined threshold); and comparing the graph coverage metric to a second predetermined threshold (corresponds to evaluation and judgement; in particular, a human, with the assistance of pen and paper, can compare the graph coverage metric to a second predetermined threshold).
Step 2A Prong Two Analysis: The judicial exceptions are not integrated into a practical application. In particular, the claim recites additional elements that are mere instructions to apply (See MPEP 2106.05(f)) or insignificant extra-solution activity (See MPEP 2106.05(g)). The limitation:
“the controller”
As drafted, is an additional element that amounts to no more than mere instructions to apply the exception for the abstract ideas. See MPEP 2106.05(f). The limitations:
“in response to the rules recovered metric being greater than or equal to the first predetermined threshold and the graph coverage metric being greater than or equal to the second predetermined threshold, save the trained link prediction model in the memory of the Al device for deployment or transmit the trained link prediction model to an external device for deployment”
As drafted, are additional elements that correspond to insignificant extra-solution activity. In particular, the additional elements are merely directed towards mere data gathering. See MPEP 2106.05(g). In addition, the recitation of additional elements in claim 14 of a generic memory, controller, and generic model training, as drafted, are reciting mere instructions to apply language such that it amounts to no more than mere instructions to apply the exceptions. Furthermore, the “obtain …” and “output …” limitations of claim 14 are additional elements that correspond to insignificant extra-solution activity as mere data gathering. Therefore, the additional elements do not integrate the abstract ideas into a practical application.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to integration of the abstract idea into a practical application, all of the additional elements are “mere instructions to apply an exception” (I.e. the additional elements describe a generic memory, controller, and generic model training for applying the abstract ideas) or insignificant extra-solution activity (i.e. obtaining/receiving, outputting/transmitting, and saving/storing data). Furthermore, the “obtain …”, “output …”, “… save …”, and “… transmit …” limitations are insignificant extra-solution activity that is well-understood, routine, and conventional according to MPEP 2106.05(d) (“The courts have recognized the following computer functions as well‐understood, routine, and conventional functions when they are claimed in a merely generic manner (e.g., at a high level of generality) or as insignificant extra-solution activity… i. Receiving or transmitting data over a network … iv. Storing and retrieving information in memory). Mere instructions to apply an exception cannot provide an inventive concept. The claim is not patent eligible.
Regarding Claim 16,
Claim 16 is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 16 is directed to a device, which is directed to a machine, one of the statutory categories.
Step 2A Prong One Analysis: The limitation:
“generate one quality score for the logic rules based on a sum of the plurality of scores and based on dividing by a total number of the logic rules”
As drafted, under their broadest reasonable interpretations, cover mental processes (concepts performed in the human mind (including an observation, evaluation, judgement, opinion)) but for the recitation of mere instructions to apply language (See MPEP 2106.05(f)) and insignificant extra-solution activity (See MPEP 2106.05(g)). The above limitations in the context of this claim encompass generating a quality score for the logic rules based on a sum of the plurality of scores based on dividing by a total number of the logic rules (corresponds to evaluation and judgement; in particular, a human, with the assistance of pen and paper, can generate one quality score for the logic rules based on a sum of the plurality of quality scores and based on dividing by a total number of the logic rules).
Step 2A Prong Two Analysis: The judicial exceptions are not integrated into a practical application. In particular, the claim recites additional elements that are mere instructions to apply (See MPEP 2106.05(f)) or insignificant extra-solution activity (See MPEP 2106.05(g)). The limitation:
“the controller”
As drafted, is an additional element that amounts to no more than mere instructions to apply the exception for the abstract ideas. See MPEP 2106.05(f). The limitations:
“receive a plurality of scores for each logic rule among the logic rules, the plurality of scores being assigned by a plurality of annotators”
As drafted, are additional elements that correspond to insignificant extra-solution activity. In particular, the additional elements are merely directed towards mere data gathering. See MPEP 2106.05(g). In addition, the recitation of additional elements in claim 11 of a generic memory, controller, and generic model training, as drafted, are reciting mere instructions to apply language such that it amounts to no more than mere instructions to apply the exceptions. Furthermore, the “obtain …” and “output …” limitations of claim 11 are additional elements that correspond to insignificant extra-solution activity as mere data gathering. Therefore, the additional elements do not integrate the abstract ideas into a practical application.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to integration of the abstract idea into a practical application, all of the additional elements are “mere instructions to apply an exception” (I.e. the additional elements describe a generic memory, controller, and generic model training for applying the abstract ideas) or insignificant extra-solution activity (i.e. obtaining/receiving and outputting/transmitting data). Furthermore, the “obtain …”, “output …”, and “receive …” limitations are insignificant extra-solution activity that is well-understood, routine, and conventional according to MPEP 2106.05(d) (“The courts have recognized the following computer functions as well‐understood, routine, and conventional functions when they are claimed in a merely generic manner (e.g., at a high level of generality) or as insignificant extra-solution activity… i. Receiving or transmitting data over a network). Mere instructions to apply an exception cannot provide an inventive concept. The claim is not patent eligible.
Regarding Claim 17,
Claim 17 is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 17 is directed to a device, which is directed to a machine, one of the statutory categories.
Step 2A Prong One Analysis: Please see the analysis of claim 16. The limitations of claim 17 are only additional elements to the abstract ideas of claim 16.
Step 2A Prong Two Analysis: The judicial exceptions are not integrated into a practical application. In particular, the claim recites additional elements that are mere instructions to apply (See MPEP 2106.05(f)) or insignificant extra-solution activity (See MPEP 2106.05(g)). The limitation:
“wherein each of the plurality of scores is a binary value of 1 or 0”
As drafted, is part of the insignificant extra-solution activity of claim 16. The limitation of claim 17 further limits the limitation of claim 16 by further defining what the received plurality of scores comprises. In addition, the recitation of additional elements in claim 16 of a generic memory, controller, and generic model training, as drafted, are reciting mere instructions to apply language such that it amounts to no more than mere instructions to apply the exceptions. Furthermore, the “obtain …”, “output …”, and “receive …” limitations of claim 16 are additional elements that correspond to insignificant extra-solution activity as mere data gathering. Therefore, the additional elements do not integrate the abstract ideas into a practical application.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to integration of the abstract idea into a practical application, all of the additional elements are “mere instructions to apply an exception” (I.e. the additional elements describe a generic memory, controller, and generic model training for applying the abstract ideas) or insignificant extra-solution activity (i.e. obtaining/receiving and outputting/transmitting data). Furthermore, the “obtain …”, “output …”, and “receive …” limitations are insignificant extra-solution activity that is well-understood, routine, and conventional according to MPEP 2106.05(d) (“The courts have recognized the following computer functions as well‐understood, routine, and conventional functions when they are claimed in a merely generic manner (e.g., at a high level of generality) or as insignificant extra-solution activity… i. Receiving or transmitting data over a network). Mere instructions to apply an exception cannot provide an inventive concept. The claim is not patent eligible.
Regarding Claim 18,
Claim 18 is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 18 is directed to a device, which is directed to a machine, one of the statutory categories.
Step 2A Prong One Analysis: The limitation:
“compare the one quality score to a third predetermined threshold”
As drafted, under their broadest reasonable interpretations, cover mental processes (concepts performed in the human mind (including an observation, evaluation, judgement, opinion)) but for the recitation of mere instructions to apply language (See MPEP 2106.05(f)) and insignificant extra-solution activity (See MPEP 2106.05(g)). The above limitations in the context of this claim encompass comparing the one quality score to a third predetermined threshold (corresponds to evaluation and judgement; in particular, a human, with the assistance of pen and paper, can compare the one quality score to a third predetermined threshold).
Step 2A Prong Two Analysis: The judicial exceptions are not integrated into a practical application. In particular, the claim recites additional elements that are mere instructions to apply (See MPEP 2106.05(f)) or insignificant extra-solution activity (See MPEP 2106.05(g)). The limitation:
“the controller”
As drafted, is an additional element that amounts to no more than mere instructions to apply the exception for the abstract ideas. See MPEP 2106.05(f). The limitations:
“in response to the one quality score being greater than or equal to the third predetermined threshold, save the trained link prediction model in the memory of the Al device for deployment or transmit the trained link prediction model to an external device for deployment”
As drafted, are additional elements that correspond to insignificant extra-solution activity. In particular, the additional elements are merely directed towards mere data gathering. See MPEP 2106.05(g). In addition, the recitation of additional elements in claim 16 of a generic memory, controller, and generic model training, as drafted, are reciting mere instructions to apply language such that it amounts to no more than mere instructions to apply the exceptions. Furthermore, the “obtain …”, “output …”, and “receive …” limitations of claim 16 are additional elements that correspond to insignificant extra-solution activity as mere data gathering. Therefore, the additional elements do not integrate the abstract ideas into a practical application.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to integration of the abstract idea into a practical application, all of the additional elements are “mere instructions to apply an exception” (I.e. the additional elements describe a generic memory, controller, and generic model training for applying the abstract ideas) or insignificant extra-solution activity (i.e. obtaining/receiving, outputting/transmitting, and saving/storing data). Furthermore, the “obtain …”, “output …”, “receive …”, “… save …”, and “… transmit …” limitations are insignificant extra-solution activity that is well-understood, routine, and conventional according to MPEP 2106.05(d) (“The courts have recognized the following computer functions as well‐understood, routine, and conventional functions when they are claimed in a merely generic manner (e.g., at a high level of generality) or as insignificant extra-solution activity… i. Receiving or transmitting data over a network … iv. Storing and retrieving information in memory). Mere instructions to apply an exception cannot provide an inventive concept. The claim is not patent eligible.
Regarding Claim 19,
Claim 19 is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 19 is directed to a device, which is directed to a machine, one of the statutory categories.
Step 2A Prong One Analysis: The limitations:
“generate link prediction metrics based on the trained link prediction model, the link prediction metrics including at least one of mean rank (MR), mean reciprocal rank (MRR), and Hit@K”
“determine whether to deploy the trained link prediction model based on the link prediction metrics”
As drafted, under their broadest reasonable interpretations, cover mental processes (concepts performed in the human mind (including an observation, evaluation, judgement, opinion)) but for the recitation of mere instructions to apply language (See MPEP 2106.05(f)) and insignificant extra-solution activity (See MPEP 2106.05(g)). The above limitations in the context of this claim encompass generating link prediction metric based on the trained link prediction model, the link prediction metrics including at least one of MR, MMR, and Hit@K (corresponds to evaluation and judgement; in particular, a human, with the assistance of pen and paper, can generate link prediction metric including at least one of MR, MMR, and Hit@K based on the trained link prediction model); and determining whether to deploy the trained link prediction model based on the link prediction metrics (corresponds to evaluation and judgement; in particular, a human, with the assistance of pen and paper, can use the link prediction metrics to determine whether to deploy the trained link prediction model).
Step 2A Prong Two Analysis: The judicial exceptions are not integrated into a practical application. In particular, the claim recites additional elements that are mere instructions to apply (See MPEP 2106.05(f)) or insignificant extra-solution activity (See MPEP 2106.05(g)). The limitation:
“the controller”
As drafted, is an additional element that amounts to no more than mere instructions to apply the exception for the abstract ideas. See MPEP 2106.05(f). mere data gathering. See MPEP 2106.05(g). In addition, the recitation of additional elements in claim 11 of a generic memory, controller, and generic model training, as drafted, are reciting mere instructions to apply language such that it amounts to no more than mere instructions to apply the exceptions. Furthermore, the “obtain …” and “output …” limitations of claim 11 are additional elements that correspond to insignificant extra-solution activity as mere data gathering. Therefore, the additional elements do not integrate the abstract ideas into a practical application.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to integration of the abstract idea into a practical application, all of the additional elements are “mere instructions to apply an exception” (I.e. the additional elements describe a generic memory, controller, and generic model training for applying the abstract ideas) or insignificant extra-solution activity (i.e. obtaining/receiving and outputting/transmitting data). Furthermore, the “obtain …” and “output …” limitations are insignificant extra-solution activity that is well-understood, routine, and conventional according to MPEP 2106.05(d) (“The courts have recognized the following computer functions as well‐understood, routine, and conventional functions when they are claimed in a merely generic manner (e.g., at a high level of generality) or as insignificant extra-solution activity… i. Receiving or transmitting data over a network). Mere instructions to apply an exception cannot provide an inventive concept. The claim is not patent eligible.
Regarding Claim 20,
Claim 20 is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 20 is directed to a method, which is directed to a process, one of the statutory categories.
Step 2A Prong One Analysis: The limitations:
“extracting, …, logic rules from the trained link prediction model”
“generating, …, a rules recovered metric by dividing a number of the logic rules by a total number of original logic rules for the knowledge graph”
“generating, …, a graph coverage metric by dividing a number of grounded paths covered by the logic rules by a total number of ground paths for the knowledge graph”
As drafted, under their broadest reasonable interpretations, cover mental processes (concepts performed in the human mind (including an observation, evaluation, judgement, opinion)) but for the recitation of mere instructions to apply language (See MPEP 2106.05(f)) and insignificant extra-solution activity (See MPEP 2106.05(g)). The above limitations in the context of this claim encompass extracting logic rules from the trained link prediction model (corresponds to evaluation and judgement; in particular, a human, with the assistance of pen and paper, can extract logic rules from the trained link prediction model); generating a rules recovered metric that is a value based on dividing a number of the logic rules by a total number of original logic rules for the knowledge graph (corresponds to evaluation and judgement; in particular, a human, with the assistance of pen and paper, can generate a rules recovered metric by dividing a number of the logic rules by a total number of original logic rules for the knowledge graph); and generating a graph coverage metric that is a value based on dividing a number of grounded paths covered by the logic rules by a total number of ground paths for the knowledge graph (corresponds to evaluation and judgement; in particular, a human, with the assistance of pen and paper, can generate a graph coverage metric by dividing a number of grounded paths covered by the logic rules by a total number of ground paths for the knowledge graph).
Step 2A Prong Two Analysis: The judicial exceptions are not integrated into a practical application. In particular, the claim recites additional elements that are mere instructions to apply (See MPEP 2106.05(f)) or insignificant extra-solution activity (See MPEP 2106.05(g)). The limitations:
“an artificial intelligence (Al) device”
“a processor”
“training, via the processor, a link prediction model on the knowledge graph to generate a trained link prediction model”
“via the processor”
As drafted, are additional elements that amount to no more than mere instructions to apply the exception for the abstract ideas. See MPEP 2106.05(f). The limitations:
“obtaining, via a processor in the Al device, a knowledge graph”
“outputting the rules recovered metric and the graph coverage metric”
As drafted, are additional elements that correspond to insignificant extra-solution activity. In particular, the additional elements are merely directed towards mere data gathering. See MPEP 2106.05(g). Therefore, the additional elements do not integrate the abstract ideas into a practical application.
Step 2B Analysis: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to integration of the abstract idea into a practical application, all of the additional elements are “mere instructions to apply an exception” (I.e. the additional elements describe a generic AI device, processor, and generic model training for applying the abstract ideas) or insignificant extra-solution activity (i.e. obtaining/receiving and outputting/transmitting data). Furthermore, the “obtaining …” and “outputting …” limitations are insignificant extra-solution activity that is well-understood, routine, and conventional according to MPEP 2106.05(d) (“The courts have recognized the following computer functions as well‐understood, routine, and conventional functions when they are claimed in a merely generic manner (e.g., at a high level of generality) or as insignificant extra-solution activity… i. Receiving or transmitting data over a network). Mere instructions to apply an exception cannot provide an inventive concept. The claim is not patent eligible.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
This application currently names joint inventors. In considering patentability of the claims the examiner presumes that the subject matter of the various claims was commonly owned as of the effective filing date of the claimed invention(s) absent any evidence to the contrary. Applicant is advised of the obligation under 37 CFR 1.56 to point out the inventor and effective filing dates of each claim that was not commonly owned as of the effective filing date of the later invention in order for the examiner to consider the applicability of 35 U.S.C. 102(b)(2)(C) for any potential 35 U.S.C. 102(a)(2) prior art against the later invention.
Claims 1, 4, 10-11, 14, and 20 are rejected under 35 U.S.C. 103 as being unpatentable over Jiao et al. (US 2022/0067030 A1) in view of Dash et al. (US 2024/0135205 A1) and further in view of Meilicke et al. ("Anytime Bottom-Up Rule Learning for Knowledge Graph Completion").
Regarding Claim 1,
Jiao et al. teaches a method for controlling an artificial intelligence (Al) device (Fig. 7A; [0058]: "FIG. 7A depicts details of a method 700 for predicting links in a knowledge graph using a Transformer model as described in connection with the present disclosure ... The method 700 can be executed as a set of computer-executable instructions executed by a computer system and encoded or stored on a computer readable medium. Further, the method 700 can be performed by gates or circuits associated with a processor, Application Specific Integrated Circuit (ASIC), a field programmable gate array (FPGA), a system on chip (SOC), or other hardware device" teaches a method for predicting links in a knowledge graph using a transformer neural network model (e.g. artificial intelligence), the method being executed by a computer system (AI device) with a processor), the method comprising:
obtaining, via a processor in the Al device, a knowledge graph ([0085]: "some embodiments include a hierarchical transformer model (400, 500) for predicting a target entity in a knowledge graph (e.g., 100) that comprises a processor (e.g., 902) and memory (e.g., 904) storing computer-executable instructions, which when executed, cause the hierarchical transformer model (400, 500) to: receive (e.g., 708), by a first level transformer block (e.g., 450), a source entity-relation pair input (e.g., 406, 408) and a neighborhood entity-relation pair input (e.g., 418, 420, 428, 430) from the knowledge graph (e.g., 100)" teaches obtaining a knowledge graph 100 via a processor and memory);
training, via the processor, a link prediction model on the knowledge graph to generate a trained link prediction model (Fig. 8; [0062]: "FIG. 8 depict details of a method 800 for training a Transformer model to predict links in a knowledge graph in accordance with aspects of the present disclosure ... The method 800 can be executed as a set of computer-executable instructions executed by a computer system and encoded or stored on a computer readable medium. Further, the method 800 can be performed by gates or circuits associated with a processor, Application Specific Integrated Circuit (ASIC), a field programmable gate array (FPGA), a system on chip (SOC), or other hardware device. Hereinafter, the method 800 shall be explained with reference to the systems, components, modules, software, data structures, user interfaces, etc. described in conjunction with FIGS. 1-7" teaches a processor performing steps for training the transformer neural network model to predict links (link prediction model) on the knowledge graph).
Jiao et al. does not appear to explicitly teach extracting, via the processor, logic rules from the trained link prediction model; generating, via the processor, at least one evaluation metric based on the logic rules; generating, via the processor, an evaluation result based on comparing the at least one evaluation metric to a predetermined criteria; and outputting, via an output unit in the Al device, the evaluation result.
However, Dash et al. teaches extracting, via the processor, logic rules from the trained link prediction model (Fig .2; [0070]-[0071]: "As shown in FIG. 2, a linear programming based rule induction (LPRI) computing tool 200 comprises a knowledge graph (KG) processing engine 210, main control logic 220, a rule optimization engine 230, a scoring and ranking engine 240 and a downstream computing system interface 250. The rule optimization engine 230 comprises further operational elements of a rule generation engine 232 and a rule selection engine 234. The scoring and ranking engine 240 computes metrics to access how well the rules and weights chosen by the rule optimization engine 230 do when applied to a test dataset, e.g., a test portion of the input knowledge graph ... The LPRI computing tool 200 operates on data structures 260-264 to perform operations for inductively, and automatically through the machine learning logic of the illustrative embodiments, learning rules for identifying relations in knowledge graphs by building a set of rules from an initial set of rules extracted from an input knowledge graph and using linear programming techniques to evaluate the generated rules to determine whether they should be included in the final set of rules based on their effect on an objective function or cost function of the linear programming solution" teaches extracting a set of rules (logical rules) for identifying relations in knowledge graphs (link prediction) from machine learning logic (trained link prediction model)).
Jiao et al. and Dash et al. are analogous to the claimed invention because they are directed towards knowledge graph link prediction.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate extracting, via the processor, logic rules from the trained link prediction model as taught by Dash et al. to the disclosed invention of Jiao et al.
One of ordinary skill in the art would have been motivated to make this modification to "provide an automated improved computing tool and improved computing tool operations/functionality for automatically generating a set of rules and corresponding weights for identifying relations in knowledge graphs which may expose missing and/or non-represented relationships between entities of the knowledge graphs ... [and] provide such improved computing tools and functionality which have improved scaling with KG size over other approaches and provides significantly improved running time than other approaches" (Dash et al. [0105]).
Jiao et al. in view of Dash et al. does not appear to explicitly teach generating, via the processor, at least one evaluation metric based on the logic rules; generating, via the processor, an evaluation result based on comparing the at least one evaluation metric to a predetermined criteria; and outputting, via an output unit in the Al device, the evaluation result.
However, Meilicke et al. teaches generating, via the processor, at least one evaluation metric based on the logic rules (Algorithm 3; Section 2, last paragraph - Section 3, second paragraph: "The confidence of a rule is usually defined as number of bodygroundings, divided by the number of those body groundings that make the head true … Another parameter is the quality criteria Q which is used to decide whether or not a rule is stored. Q can be, for example, a threshold on the confidence. We use a sampling strategy to efficiently compute the confidences of a rule using the function score ... Within a time span (repeat-until loop) the algorithm learns as many rules as possible by iteratively sampling random paths. Once the given time span is over, the rules found within this span are evaluated. Note that R contains all rules that have been learned in the previous time spans, Rs contains all rules found in the current time span, and R's contains rules found in the current time span that have also been found in one of the previous iterations. We compute the fraction |R's|/|Rs| and if this number is above the sat parameter, we increase the path (and thus rule) length by one and continue with the overall process" teaches generating at least one evaluation metric (e.g. rule confidence) based on the logic rules);
generating, via the processor, an evaluation result based on comparing the at least one evaluation metric to a predetermined criteria (Algorithm 3; Section 2, last paragraph - Section 3, second paragraph: "The confidence of a rule is usually defined as number of bodygroundings, divided by the number of those body groundings that make the head true … Another parameter is the quality criteria Q which is used to decide whether or not a rule is stored. Q can be, for example, a threshold on the confidence. We use a sampling strategy to efficiently compute the confidences of a rule using the function score ... Within a time span (repeat-until loop) the algorithm learns as many rules as possible by iteratively sampling random paths. Once the given time span is over, the rules found within this span are evaluated. Note that R contains all rules that have been learned in the previous time spans, Rs contains all rules found in the current time span, and R's contains rules found in the current time span that have also been found in one of the previous iterations. We compute the fraction |R's|/|Rs| and if this number is above the sat parameter, we increase the path (and thus rule) length by one and continue with the overall process" teaches that the at least one evaluation metric (e.g. rule confidence) is compared to a predetermined criteria (e.g. quality criteria Q threshold) to generate an evaluation result); and
outputting, via an output unit in the Al device, the evaluation result (Algorithm 3; Section 2, last paragraph - Section 3, second paragraph: "The confidence of a rule is usually defined as number of bodygroundings, divided by the number of those body groundings that make the head true … Another parameter is the quality criteria Q which is used to decide whether or not a rule is stored. Q can be, for example, a threshold on the confidence. We use a sampling strategy to efficiently compute the confidences of a rule using the function score ... Within a time span (repeat-until loop) the algorithm learns as many rules as possible by iteratively sampling random paths. Once the given time span is over, the rules found within this span are evaluated. Note that R contains all rules that have been learned in the previous time spans, Rs contains all rules found in the current time span, and R's contains rules found in the current time span that have also been found in one of the previous iterations. We compute the fraction |R's|/|Rs| and if this number is above the sat parameter, we increase the path (and thus rule) length by one and continue with the overall process" teaches that the evaluation result is output).
Jiao et al., Dash et al., and Meilicke et al. are analogous to the claimed invention because they are directed towards knowledge graph link prediction.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate generating, via the processor, at least one evaluation metric based on the logic rules; generating, via the processor, an evaluation result based on comparing the at least one evaluation metric to a predetermined criteria; and outputting, via an output unit in the Al device, the evaluation result as taught by Meilicke et al. to the disclosed invention of Jiao et al. in view of Dash et al.
One of ordinary skill in the art would have been motivated to make this modification because "our approach is significantly faster, requires less computational resources, and yields an explanation in terms of the rules that propose a candidate" (Meilicke et al. Abstract).
Regarding Claim 4,
Jiao et al. in view of Dash et al. and further in view of Meilicke et al. teaches the method of claim 1.
In addition, Meilicke et al. further teaches wherein the generating the at least one evaluation metric includes: generating a rules recovered metric, the rules recovered metric being a value based on dividing a number of the logic rules by a total number of original logic rules for the knowledge graph (Algorithm 1; Section 3, second paragraph: "Within a time span (repeat-until loop) the algorithm learns as many rules as possible by iteratively sampling random paths. Once the given time span is over, the rules found within this span are evaluated. Note that R contains all rules that have been learned in the previous time spans, Rs contains all rules found in the current time span, and R's contains rules found in the current time span that have also been found in one of the previous iterations. We compute the fraction |R's|/|Rs| and if this number is above the sat parameter, we increase the path (and thus rule) length by one and continue with the overall process" teaches generating a rules recovered metric by dividing a number of rules found in the current and past iteration (number of recovered rules) by the number of rules in the current iteration (total number of original rules) for the knowledge graph); and
generating a graph coverage metric, the graph coverage metric being a value based on dividing a number of grounded paths covered by the logic rules by a total number of ground paths for the knowledge graph (Section 2, last paragraph: "The confidence of a rule is usually defined as number of bodygroundings, divided by the number of those body groundings that make the head true" teaches generating a confidence (graph coverage metric) for a rule based on dividing a number of bodygroundings for a rule (number of grounded paths covered by the logic rules) by the number of those body groundings that make the head true (total number of ground paths from the head of the knowledge graph)).
Jiao et al., Dash et al., and Meilicke et al. are analogous to the claimed invention because they are directed towards knowledge graph link prediction.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate wherein the generating the at least one evaluation metric includes: generating a rules recovered metric, the rules recovered metric being a value based on dividing a number of the logic rules by a total number of original logic rules for the knowledge graph; and generating a graph coverage metric, the graph coverage metric being a value based on dividing a number of grounded paths covered by the logic rules by a total number of ground paths for the knowledge graph as taught by Meilicke et al. to the disclosed invention of Jiao et al. in view of Dash et al.
One of ordinary skill in the art would have been motivated to make this modification because "our approach is significantly faster, requires less computational resources, and yields an explanation in terms of the rules that propose a candidate" (Meilicke et al. Abstract).
Regarding Claim 10,
Jiao et al. in view of Dash et al. and further in view of Meilicke et al. teaches the method of claim 1.
In addition, Jiao et al. further teaches wherein the Al device includes at least one of a smart television, a mobile phone, and a home appliance device ([0105]: "In general, any device(s) or means capable of implementing the methodology illustrated herein can be used to implement the various aspects of this disclosure. Exemplary hardware that can be used for the present disclosure includes computers, handheld devices, telephones (e.g., cellular, Internet enabled, digital, analog, hybrids, and others), and other hardware known in the art" teaches that a telephone (mobile phone) may be used for implementing the methodology of the disclosure (e.g. a telephone may be used as the AI device)).
Regarding Claim 11,
Jiao et al. teaches an artificial intelligence (AI) device for providing recommendations, the Al device comprising: a memory configured to store knowledge graph information; and a controller ([0005]: "In accordance with examples of the present disclosure, methods and systems are provided that are directed to learning graph entity representations (e.g., embeddings) using hierarchical Transformers for content recommendation" teaches a system (AI device) using a transformer neural network model (e.g. artificial intelligence) for providing recommendations. Fig. 9; [0065]: "FIG. 9 is a block diagram illustrating physical components (e.g., hardware) of a computing device 900 with which aspects of the disclosure may be practiced. The computing device components described below may be suitable for the computing devices described above. In a basic configuration, the computing device 900 may include at least one processing unit 902 (e.g., a tensor, vector, or graphics processing unit) and a system memory 904" teaches a computing device (AI device) comprising a memory for performing the methods of the disclosure. [0085]: "some embodiments include a hierarchical transformer model (400, 500) for predicting a target entity in a knowledge graph (e.g., 100) that comprises a processor (e.g., 902) and memory (e.g., 904) storing computer-executable instructions, which when executed, cause the hierarchical transformer model (400, 500) to: receive (e.g., 708), by a first level transformer block (e.g., 450), a source entity-relation pair input (e.g., 406, 408) and a neighborhood entity-relation pair input (e.g., 418, 420, 428, 430) from the knowledge graph (e.g., 100)" teaches that the memory stores knowledge graph information. Fig. 7A; [0058]: "FIG. 7A depicts details of a method 700 for predicting links in a knowledge graph using a Transformer model as described in connection with the present disclosure ... The method 700 can be executed as a set of computer-executable instructions executed by a computer system and encoded or stored on a computer readable medium. Further, the method 700 can be performed by gates or circuits associated with a processor, Application Specific Integrated Circuit (ASIC), a field programmable gate array (FPGA), a system on chip (SOC), or other hardware device" teaches a method for predicting links in a knowledge graph using a transformer neural network model (e.g. artificial intelligence), the method being executed by a computer system (AI device). [0105]: "the systems and methods of this disclosure can be implemented in conjunction with a special purpose computer, a programmed microprocessor or microcontroller" teaches that the embodied system and methods can be implemented using a microcontroller (controller)) configured to:
obtain a knowledge graph ([0085]: "some embodiments include a hierarchical transformer model (400, 500) for predicting a target entity in a knowledge graph (e.g., 100) that comprises a processor (e.g., 902) and memory (e.g., 904) storing computer-executable instructions, which when executed, cause the hierarchical transformer model (400, 500) to: receive (e.g., 708), by a first level transformer block (e.g., 450), a source entity-relation pair input (e.g., 406, 408) and a neighborhood entity-relation pair input (e.g., 418, 420, 428, 430) from the knowledge graph (e.g., 100)" teaches obtaining a knowledge graph 100 via a processor and memory),
train a link prediction model on the knowledge graph to generate a trained link prediction model (Fig. 8; [0062]: "FIG. 8 depict details of a method 800 for training a Transformer model to predict links in a knowledge graph in accordance with aspects of the present disclosure ... The method 800 can be executed as a set of computer-executable instructions executed by a computer system and encoded or stored on a computer readable medium. Further, the method 800 can be performed by gates or circuits associated with a processor, Application Specific Integrated Circuit (ASIC), a field programmable gate array (FPGA), a system on chip (SOC), or other hardware device. Hereinafter, the method 800 shall be explained with reference to the systems, components, modules, software, data structures, user interfaces, etc. described in conjunction with FIGS. 1-7" teaches a processor performing steps for training the transformer neural network model to predict links (link prediction model) on the knowledge graph).
Jiao et al. does not appear to explicitly teach extract logic rules from the trained link prediction model, generate at least one evaluation metric based on the logic rules, generate an evaluation result based on comparing the at least one evaluation metric to a predetermined criteria, and output the evaluation result.
However, Dash et al. teaches extract logic rules from the trained link prediction model (Fig .2; [0070]-[0071]: "As shown in FIG. 2, a linear programming based rule induction (LPRI) computing tool 200 comprises a knowledge graph (KG) processing engine 210, main control logic 220, a rule optimization engine 230, a scoring and ranking engine 240 and a downstream computing system interface 250. The rule optimization engine 230 comprises further operational elements of a rule generation engine 232 and a rule selection engine 234. The scoring and ranking engine 240 computes metrics to access how well the rules and weights chosen by the rule optimization engine 230 do when applied to a test dataset, e.g., a test portion of the input knowledge graph ... The LPRI computing tool 200 operates on data structures 260-264 to perform operations for inductively, and automatically through the machine learning logic of the illustrative embodiments, learning rules for identifying relations in knowledge graphs by building a set of rules from an initial set of rules extracted from an input knowledge graph and using linear programming techniques to evaluate the generated rules to determine whether they should be included in the final set of rules based on their effect on an objective function or cost function of the linear programming solution" teaches extracting a set of rules (logical rules) for identifying relations in knowledge graphs (link prediction) from machine learning logic (trained link prediction model)).
Jiao et al. and Dash et al. are analogous to the claimed invention because they are directed towards knowledge graph link prediction.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate extract logic rules from the trained link prediction model as taught by Dash et al. to the disclosed invention of Jiao et al.
One of ordinary skill in the art would have been motivated to make this modification to "provide an automated improved computing tool and improved computing tool operations/functionality for automatically generating a set of rules and corresponding weights for identifying relations in knowledge graphs which may expose missing and/or non-represented relationships between entities of the knowledge graphs ... [and] provide such improved computing tools and functionality which have improved scaling with KG size over other approaches and provides significantly improved running time than other approaches" (Dash et al. [0105]).
Jiao et al. in view of Dash et al. does not appear to explicitly teach generate at least one evaluation metric based on the logic rules, generate an evaluation result based on comparing the at least one evaluation metric to a predetermined criteria, and output the evaluation result.
However, Meilicke et al. teaches generate at least one evaluation metric based on the logic rules (Algorithm 3; Section 2, last paragraph - Section 3, second paragraph: "The confidence of a rule is usually defined as number of bodygroundings, divided by the number of those body groundings that make the head true … Another parameter is the quality criteria Q which is used to decide whether or not a rule is stored. Q can be, for example, a threshold on the confidence. We use a sampling strategy to efficiently compute the confidences of a rule using the function score ... Within a time span (repeat-until loop) the algorithm learns as many rules as possible by iteratively sampling random paths. Once the given time span is over, the rules found within this span are evaluated. Note that R contains all rules that have been learned in the previous time spans, Rs contains all rules found in the current time span, and R's contains rules found in the current time span that have also been found in one of the previous iterations. We compute the fraction |R's|/|Rs| and if this number is above the sat parameter, we increase the path (and thus rule) length by one and continue with the overall process" teaches generating at least one evaluation metric (e.g. rule confidence) based on the logic rules),
generate an evaluation result based on comparing the at least one evaluation metric to a predetermined criteria (Algorithm 3; Section 2, last paragraph - Section 3, second paragraph: "The confidence of a rule is usually defined as number of bodygroundings, divided by the number of those body groundings that make the head true … Another parameter is the quality criteria Q which is used to decide whether or not a rule is stored. Q can be, for example, a threshold on the confidence. We use a sampling strategy to efficiently compute the confidences of a rule using the function score ... Within a time span (repeat-until loop) the algorithm learns as many rules as possible by iteratively sampling random paths. Once the given time span is over, the rules found within this span are evaluated. Note that R contains all rules that have been learned in the previous time spans, Rs contains all rules found in the current time span, and R's contains rules found in the current time span that have also been found in one of the previous iterations. We compute the fraction |R's|/|Rs| and if this number is above the sat parameter, we increase the path (and thus rule) length by one and continue with the overall process" teaches that the at least one evaluation metric (e.g. rule confidence) is compared to a predetermined criteria (e.g. quality criteria Q threshold) to generate an evaluation result), and
output the evaluation result (Algorithm 3; Section 2, last paragraph - Section 3, second paragraph: "The confidence of a rule is usually defined as number of bodygroundings, divided by the number of those body groundings that make the head true … Another parameter is the quality criteria Q which is used to decide whether or not a rule is stored. Q can be, for example, a threshold on the confidence. We use a sampling strategy to efficiently compute the confidences of a rule using the function score ... Within a time span (repeat-until loop) the algorithm learns as many rules as possible by iteratively sampling random paths. Once the given time span is over, the rules found within this span are evaluated. Note that R contains all rules that have been learned in the previous time spans, Rs contains all rules found in the current time span, and R's contains rules found in the current time span that have also been found in one of the previous iterations. We compute the fraction |R's|/|Rs| and if this number is above the sat parameter, we increase the path (and thus rule) length by one and continue with the overall process" teaches that the evaluation result is output).
Jiao et al., Dash et al., and Meilicke et al. are analogous to the claimed invention because they are directed towards knowledge graph link prediction.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate generate at least one evaluation metric based on the logic rules, generate an evaluation result based on comparing the at least one evaluation metric to a predetermined criteria, and output the evaluation result as taught by Meilicke et al. to the disclosed invention of Jiao et al. in view of Dash et al.
One of ordinary skill in the art would have been motivated to make this modification because "our approach is significantly faster, requires less computational resources, and yields an explanation in terms of the rules that propose a candidate" (Meilicke et al. Abstract).
Regarding Claim 14,
Jiao et al. in view of Dash et al. and further in view of Meilicke et al. teaches the Al device of claim 11.
In addition, Meilicke et al. further teaches wherein the controller is further configured to: generate a rules recovered metric, the rules recovered metric being a value based on dividing a number of the logic rules by a total number of original logic rules for the knowledge graph (Algorithm 1; Section 3, second paragraph: "Within a time span (repeat-until loop) the algorithm learns as many rules as possible by iteratively sampling random paths. Once the given time span is over, the rules found within this span are evaluated. Note that R contains all rules that have been learned in the previous time spans, Rs contains all rules found in the current time span, and R's contains rules found in the current time span that have also been found in one of the previous iterations. We compute the fraction |R's|/|Rs| and if this number is above the sat parameter, we increase the path (and thus rule) length by one and continue with the overall process" teaches generating a rules recovered metric by dividing a number of rules found in the current and past iteration (number of recovered rules) by the number of rules in the current iteration (total number of original rules) for the knowledge graph), and
generate a graph coverage metric, the graph coverage metric being a value based on dividing a number of grounded paths covered by the logic rules by a total number of ground paths for the knowledge graph (Section 2, last paragraph: "The confidence of a rule is usually defined as number of bodygroundings, divided by the number of those body groundings that make the head true" teaches generating a confidence (graph coverage metric) for a rule based on dividing a number of bodygroundings for a rule (number of grounded paths covered by the logic rules) by the number of those body groundings that make the head true (total number of ground paths from the head of the knowledge graph)).
Jiao et al., Dash et al., and Meilicke et al. are analogous to the claimed invention because they are directed towards knowledge graph link prediction.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate wherein the controller is further configured to: generate a rules recovered metric, the rules recovered metric being a value based on dividing a number of the logic rules by a total number of original logic rules for the knowledge graph, and generate a graph coverage metric, the graph coverage metric being a value based on dividing a number of grounded paths covered by the logic rules by a total number of ground paths for the knowledge graph as taught by Meilicke et al. to the disclosed invention of Jiao et al. in view of Dash et al.
One of ordinary skill in the art would have been motivated to make this modification because "our approach is significantly faster, requires less computational resources, and yields an explanation in terms of the rules that propose a candidate" (Meilicke et al. Abstract).
Regarding Claim 20,
Jiao et al. teaches a method for controlling an artificial intelligence (Al) device (Fig. 7A; [0058]: "FIG. 7A depicts details of a method 700 for predicting links in a knowledge graph using a Transformer model as described in connection with the present disclosure ... The method 700 can be executed as a set of computer-executable instructions executed by a computer system and encoded or stored on a computer readable medium. Further, the method 700 can be performed by gates or circuits associated with a processor, Application Specific Integrated Circuit (ASIC), a field programmable gate array (FPGA), a system on chip (SOC), or other hardware device" teaches a method for predicting links in a knowledge graph using a transformer neural network model (e.g. artificial intelligence), the method being executed by a computer system (AI device) with a processor), the method comprising:
obtaining, via a processor in the Al device, a knowledge graph ([0085]: "some embodiments include a hierarchical transformer model (400, 500) for predicting a target entity in a knowledge graph (e.g., 100) that comprises a processor (e.g., 902) and memory (e.g., 904) storing computer-executable instructions, which when executed, cause the hierarchical transformer model (400, 500) to: receive (e.g., 708), by a first level transformer block (e.g., 450), a source entity-relation pair input (e.g., 406, 408) and a neighborhood entity-relation pair input (e.g., 418, 420, 428, 430) from the knowledge graph (e.g., 100)" teaches obtaining a knowledge graph 100 via a processor and memory);
training, via the processor, a link prediction model on the knowledge graph to generate a trained link prediction model (Fig. 8; [0062]: "FIG. 8 depict details of a method 800 for training a Transformer model to predict links in a knowledge graph in accordance with aspects of the present disclosure ... The method 800 can be executed as a set of computer-executable instructions executed by a computer system and encoded or stored on a computer readable medium. Further, the method 800 can be performed by gates or circuits associated with a processor, Application Specific Integrated Circuit (ASIC), a field programmable gate array (FPGA), a system on chip (SOC), or other hardware device. Hereinafter, the method 800 shall be explained with reference to the systems, components, modules, software, data structures, user interfaces, etc. described in conjunction with FIGS. 1-7" teaches a processor performing steps for training the transformer neural network model to predict links (link prediction model) on the knowledge graph).
Jiao et al. does not appear to explicitly teach extracting, via the processor, logic rules from the trained link prediction model; generating, via the processor, a rules recovered metric by dividing a number of the logic rules by a total number of original logic rules for the knowledge graph; generating, via the processor, a graph coverage metric by dividing a number of grounded paths covered by the logic rules by a total number of ground paths for the knowledge graph; and outputting the rules recovered metric and the graph coverage metric.
However, Dash et al. teaches extracting, via the processor, logic rules from the trained link prediction model (Fig .2; [0070]-[0071]: "As shown in FIG. 2, a linear programming based rule induction (LPRI) computing tool 200 comprises a knowledge graph (KG) processing engine 210, main control logic 220, a rule optimization engine 230, a scoring and ranking engine 240 and a downstream computing system interface 250. The rule optimization engine 230 comprises further operational elements of a rule generation engine 232 and a rule selection engine 234. The scoring and ranking engine 240 computes metrics to access how well the rules and weights chosen by the rule optimization engine 230 do when applied to a test dataset, e.g., a test portion of the input knowledge graph ... The LPRI computing tool 200 operates on data structures 260-264 to perform operations for inductively, and automatically through the machine learning logic of the illustrative embodiments, learning rules for identifying relations in knowledge graphs by building a set of rules from an initial set of rules extracted from an input knowledge graph and using linear programming techniques to evaluate the generated rules to determine whether they should be included in the final set of rules based on their effect on an objective function or cost function of the linear programming solution" teaches extracting a set of rules (logical rules) for identifying relations in knowledge graphs (link prediction) from machine learning logic (trained link prediction model)).
Jiao et al. and Dash et al. are analogous to the claimed invention because they are directed towards knowledge graph link prediction.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate extracting, via the processor, logic rules from the trained link prediction model as taught by Dash et al. to the disclosed invention of Jiao et al.
One of ordinary skill in the art would have been motivated to make this modification to "provide an automated improved computing tool and improved computing tool operations/functionality for automatically generating a set of rules and corresponding weights for identifying relations in knowledge graphs which may expose missing and/or non-represented relationships between entities of the knowledge graphs ... [and] provide such improved computing tools and functionality which have improved scaling with KG size over other approaches and provides significantly improved running time than other approaches" (Dash et al. [0105]).
Jiao et al. in view of Dash et al. does not appear to explicitly teach generating, via the processor, a rules recovered metric by dividing a number of the logic rules by a total number of original logic rules for the knowledge graph; generating, via the processor, a graph coverage metric by dividing a number of grounded paths covered by the logic rules by a total number of ground paths for the knowledge graph; and outputting the rules recovered metric and the graph coverage metric.
However, Meilicke et al. teaches generating, via the processor, a rules recovered metric by dividing a number of the logic rules by a total number of original logic rules for the knowledge graph (Algorithm 1; Section 3, second paragraph: "Within a time span (repeat-until loop) the algorithm learns as many rules as possible by iteratively sampling random paths. Once the given time span is over, the rules found within this span are evaluated. Note that R contains all rules that have been learned in the previous time spans, Rs contains all rules found in the current time span, and R's contains rules found in the current time span that have also been found in one of the previous iterations. We compute the fraction |R's|/|Rs| and if this number is above the sat parameter, we increase the path (and thus rule) length by one and continue with the overall process" teaches generating a rules recovered metric by dividing a number of rules found in the current and past iteration (number of recovered rules) by the number of rules in the current iteration (total number of original rules) for the knowledge graph);
generating, via the processor, a graph coverage metric by dividing a number of grounded paths covered by the logic rules by a total number of ground paths for the knowledge graph (Section 2, last paragraph: "The confidence of a rule is usually defined as number of bodygroundings, divided by the number of those body groundings that make the head true" teaches generating a confidence (graph coverage metric) for a rule based on dividing a number of bodygroundings for a rule (number of grounded paths covered by the logic rules) by the number of those body groundings that make the head true (total number of ground paths from the head of the knowledge graph)); and
outputting the rules recovered metric and the graph coverage metric (Algorithm 3; Section 2, last paragraph - Section 3, second paragraph: "The confidence of a rule is usually defined as number of bodygroundings, divided by the number of those body groundings that make the head true … Another parameter is the quality criteria Q which is used to decide whether or not a rule is stored. Q can be, for example, a threshold on the confidence. We use a sampling strategy to efficiently compute the confidences of a rule using the function score ... Within a time span (repeat-until loop) the algorithm learns as many rules as possible by iteratively sampling random paths. Once the given time span is over, the rules found within this span are evaluated. Note that R contains all rules that have been learned in the previous time spans, Rs contains all rules found in the current time span, and R's contains rules found in the current time span that have also been found in one of the previous iterations. We compute the fraction |R's|/|Rs| and if this number is above the sat parameter, we increase the path (and thus rule) length by one and continue with the overall process" teaches that the rules recovered and graph coverage metrics are output).
Jiao et al., Dash et al., and Meilicke et al. are analogous to the claimed invention because they are directed towards knowledge graph link prediction.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate generating, via the processor, a rules recovered metric by dividing a number of the logic rules by a total number of original logic rules for the knowledge graph; generating, via the processor, a graph coverage metric by dividing a number of grounded paths covered by the logic rules by a total number of ground paths for the knowledge graph; and outputting the rules recovered metric and the graph coverage metric as taught by Meilicke et al. to the disclosed invention of Jiao et al. in view of Dash et al.
One of ordinary skill in the art would have been motivated to make this modification because "our approach is significantly faster, requires less computational resources, and yields an explanation in terms of the rules that propose a candidate" (Meilicke et al. Abstract).
Claims 2-3, 5-8, 12-13, and 15-18 are rejected under 35 U.S.C. 103 as being unpatentable over Jiao et al. (US 2022/0067030 A1) in view of Dash et al. (US 2024/0135205 A1) in view of Meilicke et al. ("Anytime Bottom-Up Rule Learning for Knowledge Graph Completion") and further in view of Gu ("Learning Logical Rules from Knowledge Graphs").
Regarding Claim 2,
Jiao et al. in view of Dash et al. and further in view of Meilicke et al. teaches the method of claim 1.
Jiao et al. in view of Dash et al. and further in view of Meilicke et al. does not appear to explicitly teach further comprising: saving the trained link prediction model in a memory of the Al device for deployment or transmitting the trained link prediction model to an external device for deployment, based on the evaluation results.
However, Gu teaches further comprising: saving the trained link prediction model in a memory of the Al device for deployment or transmitting the trained link prediction model to an external device for deployment, based on the evaluation results (Section 2.2, last paragraph: "given a target predicate r, a background knowledge B and sets of positive and negative instances of r, denoted by I+ and I− respectively, a rule learning algorithm aims to produce a set of rules P. When rules in P are grounded over B, CP(I+) and CP(I−) are satisfactory to specific criteria. The learned model is thus the rule set P" teaches that when the rule set determined by the rule learning algorithm is satisfactory to specific criteria (e.g. based on the evaluation results), the learned model is then saved as the rule set (e.g. trained model is saved in memory of AI device)).
Jiao et al., Dash et al., Meilicke et al., and Gu are analogous to the claimed invention because they are directed towards knowledge graph link prediction.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate further comprising: saving the trained link prediction model in a memory of the Al device for deployment or transmitting the trained link prediction model to an external device for deployment, based on the evaluation results as taught by Gu to the disclosed invention of Jiao et al. in view of Dash et al. and further in view of Meilicke et al.
One of ordinary skill in the art would have been motivated to make this modification to "significantly reduces the runtime on evaluating instantiated rules, discovers much more high-quality rules than existing works and performs competitively on knowledge graph completion task compared to existing methods" (Gu Section 6, first paragraph).
Regarding Claim 3,
Jiao et al. in view of Dash et al. in view of Meilicke et al. and further in view of Gu teaches the method of claim 2.
In addition, Jiao et al. further teaches wherein the trained link prediction model is deployed in a question and answer system or a recommendation system (Fig. 3; [0041]: "FIG. 3 illustrates a context-independent transformer model 300 that may be used to predict a link (e.g., a target node) in a knowledge graph, such as knowledge graph 100. The model 300 may be used to fill in a missing target node in a knowledge graph or to respond to a query for content or as part of a content recommendation system" teaches that the trained link prediction model is used (deployed) in a query answering (question and answer) or content recommendation system).
Regarding Claim 5,
Jiao et al. in view of Dash et al. and further in view of Meilicke et al. teaches the method of claim 4.
In addition, Meilicke et al. further teaches further comprising: comparing the rules recovered metric to a first predetermined threshold (Algorithm 1; Section 3, second paragraph: "within a time span (repeat-until loop) the algorithm learns as many rules as possible by iteratively sampling random paths. Once the given time span is over, the rules found within this span are evaluated. Note that R contains all rules that have been learned in the previous time spans, Rs contains all rules found in the current time span, and R's contains rules found in the current time span that have also been found in one of the previous iterations. We compute the fraction |R's|/|Rs| and if this number is above the sat parameter, we increase the path (and thus rule) length by one and continue with the overall process" teaches generating a rules recovered metric by dividing a number of rules found in the current and past iteration (number of recovered rules) by the number of rules in the current iteration (total number of original rules) for the knowledge graph and comparing it to a sat parameter (first predetermined threshold));
comparing the graph coverage metric to a second predetermined threshold (Algorithm 1; Section 3, first paragraph: "Another parameter is the quality criteria Q which is used to decide whether or not a rule is stored. Q can be, for example, a threshold on the confidence. We use a sampling strategy to efficiently compute the confidences of a rule using the function score" teaches that the confidence of the rule (graph coverage metric) is compared to a threshold quality criteria Q (second predetermined threshold)).
Jiao et al., Dash et al., and Meilicke et al. are analogous to the claimed invention because they are directed towards knowledge graph link prediction.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate further comprising: comparing the rules recovered metric to a first predetermined threshold; comparing the graph coverage metric to a second predetermined threshold as taught by Meilicke et al. to the disclosed invention of Jiao et al. in view of Dash et al.
One of ordinary skill in the art would have been motivated to make this modification because "our approach is significantly faster, requires less computational resources, and yields an explanation in terms of the rules that propose a candidate" (Meilicke et al. Abstract).
Jiao et al. in view of Dash et al. and further in view of Meilicke et al. does not appear to explicitly teach in response to the rules recovered metric being greater than or equal to the first predetermined threshold and the graph coverage metric being greater than or equal to the second predetermined threshold, saving the trained link prediction model in a memory of the Al device for deployment or transmitting the trained link prediction model to an external device for deployment.
However, Gu teaches in response to the rules recovered metric being greater than or equal to the first predetermined threshold and the graph coverage metric being greater than or equal to the second predetermined threshold, saving the trained link prediction model in a memory of the Al device for deployment or transmitting the trained link prediction model to an external device for deployment (Section 2.2, last paragraph: "given a target predicate r, a background knowledge B and sets of positive and negative instances of r, denoted by I+ and I− respectively, a rule learning algorithm aims to produce a set of rules P. When rules in P are grounded over B, CP(I+) and CP(I−) are satisfactory to specific criteria. The learned model is thus the rule set P" teaches that when the rule set determined by the rule learning algorithm is satisfactory to specific criteria (e.g. the metrics are greater than or equal to the first and second predetermined thresholds), the learned model is then saved as the rule set (e.g. trained model is saved in memory of AI device)).
Jiao et al., Dash et al., Meilicke et al., and Gu are analogous to the claimed invention because they are directed towards knowledge graph link prediction.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate in response to the rules recovered metric being greater than or equal to the first predetermined threshold and the graph coverage metric being greater than or equal to the second predetermined threshold, saving the trained link prediction model in a memory of the Al device for deployment or transmitting the trained link prediction model to an external device for deployment as taught by Gu to the disclosed invention of Jiao et al. in view of Dash et al. and further in view of Meilicke et al.
One of ordinary skill in the art would have been motivated to make this modification to "significantly reduces the runtime on evaluating instantiated rules, discovers much more high-quality rules than existing works and performs competitively on knowledge graph completion task compared to existing methods" (Gu Section 6, first paragraph).
Regarding Claim 6,
Jiao et al. in view of Dash et al. and further in view of Meilicke et al. teaches the method of claim 1.
Jiao et al. in view of Dash et al. and further in view of Meilicke et al. does not appear to explicitly teach wherein the generating the at least one evaluation metric includes: receiving a plurality of scores for each logic rule among the logic rules, the plurality of scores being assigned by a plurality of annotators; and generating one quality score for the logic rules based on a sum of the plurality of scores and based on dividing by a total number of the logic rules.
However, Gu teaches wherein the generating the at least one evaluation metric includes: receiving a plurality of scores for each logic rule among the logic rules, the plurality of scores being assigned by a plurality of annotators (Equation 3.2; Section 3.2.3, second paragraph: "For each abstract rule p ∈ P, GPFL grounds it over G to produce groundings G. Groundings can be used to score the abstract rule p if p is a CAR, or to derive and evaluate instantiated rules. We define a scoring procedure Score that first measures the quality of a rule and then decides if the rule is good enough to be included in the rule set. Various rule quality measures have been proposed in existing works. In this work, we employ three popular measures, namely the standard confidence [39], the smooth confidence [75] and the Partial Completeness Assumption (PCA) [39]. First, we define the support of a rule p as:
PNG
media_image1.png
42
384
media_image1.png
Greyscale
which counts the number of positive instances covered by the rule" teaches receiving quality measures (plurality of scores) for each rule (logic rule) among the rules, the quality measures being assigned by a scoring procedure (e.g. using annotators)); and
generating one quality score for the logic rules based on a sum of the plurality of scores and based on dividing by a total number of the logic rules (Equation 3.2; Section 3.2.3, second paragraph: "For each abstract rule p ∈ P, GPFL grounds it over G to produce groundings G. Groundings can be used to score the abstract rule p if p is a CAR, or to derive and evaluate instantiated rules. We define a scoring procedure Score that first measures the quality of a rule and then decides if the rule is good enough to be included in the rule set. Various rule quality measures have been proposed in existing works. In this work, we employ three popular measures, namely the standard confidence [39], the smooth confidence [75] and the Partial Completeness Assumption (PCA) [39]. First, we define the support of a rule p as:
PNG
media_image1.png
42
384
media_image1.png
Greyscale
which counts the number of positive instances covered by the rule" teaches receiving quality measures (plurality of scores) for each rule (logic rule) among the rules. Equation 3.8; Equation 3.9; Section 3.3.5, first paragraph: "we first define metrics indicating the predictive performance of learned rules . The test precision of a rule p is defined as:
PNG
media_image2.png
52
396
media_image2.png
Greyscale
where suppt(p)=|Hp∩It+|.We propose and use the Global Average Precision (GAP), the average of the average test precision of top-k rules of each target over all targets, as the performance indicator, where the rules are sorted by a quality measure. Formally, given a set of target predicates R, we define GAP as:
PNG
media_image3.png
66
406
media_image3.png
Greyscale
where Prt contains the top-k rules with the target predicate rt. Similarly, we also measure the Global Average Quality (GAQ) over target predicates by replacing the precision function in Equation.3.9 with a quality measure function" teaches generating a global average quality (GAQ) (one quality score) for the rules (logic rules) based on a sum of the quality measures for the rules divided by the total number of rules).
Jiao et al., Dash et al., Meilicke et al., and Gu are analogous to the claimed invention because they are directed towards knowledge graph link prediction.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate wherein the generating the at least one evaluation metric includes: receiving a plurality of scores for each logic rule among the logic rules, the plurality of scores being assigned by a plurality of annotators; and generating one quality score for the logic rules based on a sum of the plurality of scores and based on dividing by a total number of the logic rules as taught by Gu to the disclosed invention of Jiao et al. in view of Dash et al. and further in view of Meilicke et al.
One of ordinary skill in the art would have been motivated to make this modification to "significantly reduces the runtime on evaluating instantiated rules, discovers much more high-quality rules than existing works and performs competitively on knowledge graph completion task compared to existing methods" (Gu Section 6, first paragraph).
Regarding Claim 7,
Jiao et al. in view of Dash et al. in view of Meilicke et al. and further in view of Gu teaches the method of claim 6.
In addition, Gu further teaches wherein each of the plurality of scores is a binary value of 1 or 0 (Equation 2.4; Section 2.2.2, first paragraph: "a comprehensive solution set contains all the qualified rules under certain local criteria. We denote by φ(P′) an indicator function that returns true if for any p ∈ P′, p is considered satisfactory to a collection of pre-defined local criteria, e.g., the coverage of positive instances or prediction confidence" teaches that quality scores for the rules are determined based on an indicator function (e.g. outputs binary score of 0 or 1) to determine if each rule is a qualified rule or not).
Jiao et al., Dash et al., Meilicke et al., and Gu are analogous to the claimed invention because they are directed towards knowledge graph link prediction.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate wherein each of the plurality of scores is a binary value of 1 or 0 as taught by Gu to the disclosed invention of Jiao et al. in view of Dash et al. and further in view of Meilicke et al.
One of ordinary skill in the art would have been motivated to make this modification because "a comprehensive solution set provides better predictive performance and interpretability than a concise set at the expense of model size and management overhead. The improvements in performance are mainly attributed to the fact that the rules in a comprehensive set can provide a more complete knowledge about target concepts" (Gu Section 2.2.2, first paragraph).
Regarding Claim 8,
Jiao et al. in view of Dash et al. in view of Meilicke et al. and further in view of Gu teaches the method of claim 6.
In addition, Gu further teaches further comprising: comparing the one quality score to a third predetermined threshold (Section 2.2.2, first paragraph: "a comprehensive solution set contains all the qualified rules under certain local criteria. We denote by φ(P′) an indicator function that returns true if for any p ∈ P′, p is considered satisfactory to a collection of pre-defined local criteria, e.g., the coverage of positive instances or prediction confidence" teaches comparing the satisfactory rule set from the learning system (one quality score) to some criteria (third predetermined threshold)); and
in response to the one quality score being greater than or equal to the third predetermined threshold, saving the trained link prediction model in a memory of the Al device for deployment or transmitting the trained link prediction model to an external device for deployment (Section 2.2, last paragraph: "given a target predicate r, a background knowledge B and sets of positive and negative instances of r, denoted by I+ and I− respectively, a rule learning algorithm aims to produce a set of rules P. When rules in P are grounded over B, CP(I+) and CP(I−) are satisfactory to specific criteria. The learned model is thus the rule set P" teaches that when the rule set determined by the rule learning algorithm is satisfactory to specific criteria (e.g. the one quality score is greater than or equal to a third predetermined threshold), the learned model is then saved as the rule set (e.g. trained model is saved in memory of AI device)).
Jiao et al., Dash et al., Meilicke et al., and Gu are analogous to the claimed invention because they are directed towards knowledge graph link prediction.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate further comprising: comparing the one quality score to a third predetermined threshold; and in response to the one quality score being greater than or equal to the third predetermined threshold, saving the trained link prediction model in a memory of the Al device for deployment or transmitting the trained link prediction model to an external device for deployment as taught by Gu to the disclosed invention of Jiao et al. in view of Dash et al. and further in view of Meilicke et al.
One of ordinary skill in the art would have been motivated to make this modification to "significantly reduces the runtime on evaluating instantiated rules, discovers much more high-quality rules than existing works and performs competitively on knowledge graph completion task compared to existing methods" (Gu Section 6, first paragraph).
Regarding Claim 12,
Jiao et al. in view of Dash et al. and further in view of Meilicke et al. teaches the AI device of claim 11.
Jiao et al. in view of Dash et al. and further in view of Meilicke et al. does not appear to explicitly teach wherein the controller is further configured to: save the trained link prediction model in the memory of the Al device for deployment or transmit the trained link prediction model to an external device for deployment, based on the evaluation results.
However, Gu teaches wherein the controller is further configured to: save the trained link prediction model in the memory of the Al device for deployment or transmit the trained link prediction model to an external device for deployment, based on the evaluation results (Section 2.2, last paragraph: "given a target predicate r, a background knowledge B and sets of positive and negative instances of r, denoted by I+ and I− respectively, a rule learning algorithm aims to produce a set of rules P. When rules in P are grounded over B, CP(I+) and CP(I−) are satisfactory to specific criteria. The learned model is thus the rule set P" teaches that when the rule set determined by the rule learning algorithm is satisfactory to specific criteria (e.g. based on the evaluation results), the learned model is then saved as the rule set (e.g. trained model is saved in memory of AI device)).
Jiao et al., Dash et al., Meilicke et al., and Gu are analogous to the claimed invention because they are directed towards knowledge graph link prediction.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate wherein the controller is further configured to: save the trained link prediction model in the memory of the Al device for deployment or transmit the trained link prediction model to an external device for deployment, based on the evaluation results as taught by Gu to the disclosed invention of Jiao et al. in view of Dash et al. and further in view of Meilicke et al.
One of ordinary skill in the art would have been motivated to make this modification to "significantly reduces the runtime on evaluating instantiated rules, discovers much more high-quality rules than existing works and performs competitively on knowledge graph completion task compared to existing methods" (Gu Section 6, first paragraph).
Regarding Claim 13,
Jiao et al. in view of Dash et al. in view of Meilicke et al. and further in view of Gu teaches the AI device of claim 12.
In addition, Jiao et al. further teaches wherein the trained link prediction model is deployed in a question and answer system or a recommendation system (Fig. 3; [0041]: "FIG. 3 illustrates a context-independent transformer model 300 that may be used to predict a link (e.g., a target node) in a knowledge graph, such as knowledge graph 100. The model 300 may be used to fill in a missing target node in a knowledge graph or to respond to a query for content or as part of a content recommendation system" teaches that the trained link prediction model is used (deployed) in a query answering (question and answer) or content recommendation system).
Regarding Claim 15,
Jiao et al. in view of Dash et al. and further in view of Meilicke et al. teaches the AI device of claim 14.
In addition, Meilicke et al. further teaches wherein the controller is further configured to: compare the rules recovered metric to a first predetermined threshold (Algorithm 1; Section 3, second paragraph: "within a time span (repeat-until loop) the algorithm learns as many rules as possible by iteratively sampling random paths. Once the given time span is over, the rules found within this span are evaluated. Note that R contains all rules that have been learned in the previous time spans, Rs contains all rules found in the current time span, and R's contains rules found in the current time span that have also been found in one of the previous iterations. We compute the fraction |R's|/|Rs| and if this number is above the sat parameter, we increase the path (and thus rule) length by one and continue with the overall process" teaches generating a rules recovered metric by dividing a number of rules found in the current and past iteration (number of recovered rules) by the number of rules in the current iteration (total number of original rules) for the knowledge graph and comparing it to a sat parameter (first predetermined threshold)),
compare the graph coverage metric to a second predetermined threshold (Algorithm 1; Section 3, first paragraph: "Another parameter is the quality criteria Q which is used to decide whether or not a rule is stored. Q can be, for example, a threshold on the confidence. We use a sampling strategy to efficiently compute the confidences of a rule using the function score" teaches that the confidence of the rule (graph coverage metric) is compared to a threshold quality criteria Q (second predetermined threshold)).
Jiao et al., Dash et al., and Meilicke et al. are analogous to the claimed invention because they are directed towards knowledge graph link prediction.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate wherein the controller is further configured to: compare the rules recovered metric to a first predetermined threshold, compare the graph coverage metric to a second predetermined threshold as taught by Meilicke et al. to the disclosed invention of Jiao et al. in view of Dash et al.
One of ordinary skill in the art would have been motivated to make this modification because "our approach is significantly faster, requires less computational resources, and yields an explanation in terms of the rules that propose a candidate" (Meilicke et al. Abstract).
Jiao et al. in view of Dash et al. and further in view of Meilicke et al. does not appear to explicitly teach in response to the rules recovered metric being greater than or equal to the first predetermined threshold and the graph coverage metric being greater than or equal to the second predetermined threshold, save the trained link prediction model in the memory of the Al device for deployment or transmit the trained link prediction model to an external device for deployment.
However, Gu teaches in response to the rules recovered metric being greater than or equal to the first predetermined threshold and the graph coverage metric being greater than or equal to the second predetermined threshold, save the trained link prediction model in the memory of the Al device for deployment or transmit the trained link prediction model to an external device for deployment (Section 2.2, last paragraph: "given a target predicate r, a background knowledge B and sets of positive and negative instances of r, denoted by I+ and I− respectively, a rule learning algorithm aims to produce a set of rules P. When rules in P are grounded over B, CP(I+) and CP(I−) are satisfactory to specific criteria. The learned model is thus the rule set P" teaches that when the rule set determined by the rule learning algorithm is satisfactory to specific criteria (e.g. the metrics are greater than or equal to the first and second predetermined thresholds), the learned model is then saved as the rule set (e.g. trained model is saved in memory of AI device)).
Jiao et al., Dash et al., Meilicke et al., and Gu are analogous to the claimed invention because they are directed towards knowledge graph link prediction.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate in response to the rules recovered metric being greater than or equal to the first predetermined threshold and the graph coverage metric being greater than or equal to the second predetermined threshold, save the trained link prediction model in the memory of the Al device for deployment or transmit the trained link prediction model to an external device for deployment as taught by Gu to the disclosed invention of Jiao et al. in view of Dash et al. and further in view of Meilicke et al.
One of ordinary skill in the art would have been motivated to make this modification to "significantly reduces the runtime on evaluating instantiated rules, discovers much more high-quality rules than existing works and performs competitively on knowledge graph completion task compared to existing methods" (Gu Section 6, first paragraph).
Regarding Claim 16,
Jiao et al. in view of Dash et al. and further in view of Meilicke et al. teaches the AI device of claim 11.
Jiao et al. in view of Dash et al. and further in view of Meilicke et al. does not appear to explicitly teach wherein the controller is further configured to: receive a plurality of scores for each logic rule among the logic rules, the plurality of scores being assigned by a plurality of annotators, and generate one quality score for the logic rules based on a sum of the plurality of scores and based on dividing by a total number of the logic rules.
However, Gu teaches wherein the controller is further configured to: receive a plurality of scores for each logic rule among the logic rules, the plurality of scores being assigned by a plurality of annotators (Equation 3.2; Section 3.2.3, second paragraph: "For each abstract rule p ∈ P, GPFL grounds it over G to produce groundings G. Groundings can be used to score the abstract rule p if p is a CAR, or to derive and evaluate instantiated rules. We define a scoring procedure Score that first measures the quality of a rule and then decides if the rule is good enough to be included in the rule set. Various rule quality measures have been proposed in existing works. In this work, we employ three popular measures, namely the standard confidence [39], the smooth confidence [75] and the Partial Completeness Assumption (PCA) [39]. First, we define the support of a rule p as:
PNG
media_image1.png
42
384
media_image1.png
Greyscale
which counts the number of positive instances covered by the rule" teaches receiving quality measures (plurality of scores) for each rule (logic rule) among the rules, the quality measures being assigned by a scoring procedure (e.g. using annotators)), and
generate one quality score for the logic rules based on a sum of the plurality of scores and based on dividing by a total number of the logic rules (Equation 3.2; Section 3.2.3, second paragraph: "For each abstract rule p ∈ P, GPFL grounds it over G to produce groundings G. Groundings can be used to score the abstract rule p if p is a CAR, or to derive and evaluate instantiated rules. We define a scoring procedure Score that first measures the quality of a rule and then decides if the rule is good enough to be included in the rule set. Various rule quality measures have been proposed in existing works. In this work, we employ three popular measures, namely the standard confidence [39], the smooth confidence [75] and the Partial Completeness Assumption (PCA) [39]. First, we define the support of a rule p as:
PNG
media_image1.png
42
384
media_image1.png
Greyscale
which counts the number of positive instances covered by the rule" teaches receiving quality measures (plurality of scores) for each rule (logic rule) among the rules. Equation 3.8; Equation 3.9; Section 3.3.5, first paragraph: "we first define metrics indicating the predictive performance of learned rules . The test precision of a rule p is defined as:
PNG
media_image2.png
52
396
media_image2.png
Greyscale
where suppt(p)=|Hp∩It+|.We propose and use the Global Average Precision (GAP), the average of the average test precision of top-k rules of each target over all targets, as the performance indicator, where the rules are sorted by a quality measure. Formally, given a set of target predicates R, we define GAP as:
PNG
media_image3.png
66
406
media_image3.png
Greyscale
where Prt contains the top-k rules with the target predicate rt. Similarly, we also measure the Global Average Quality (GAQ) over target predicates by replacing the precision function in Equation.3.9 with a quality measure function" teaches generating a global average quality (GAQ) (one quality score) for the rules (logic rules) based on a sum of the quality measures for the rules divided by the total number of rules).
Jiao et al., Dash et al., Meilicke et al., and Gu are analogous to the claimed invention because they are directed towards knowledge graph link prediction.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate wherein the controller is further configured to: receive a plurality of scores for each logic rule among the logic rules, the plurality of scores being assigned by a plurality of annotators, and generate one quality score for the logic rules based on a sum of the plurality of scores and based on dividing by a total number of the logic rules as taught by Gu to the disclosed invention of Jiao et al. in view of Dash et al. and further in view of Meilicke et al.
One of ordinary skill in the art would have been motivated to make this modification to "significantly reduces the runtime on evaluating instantiated rules, discovers much more high-quality rules than existing works and performs competitively on knowledge graph completion task compared to existing methods" (Gu Section 6, first paragraph).
Regarding Claim 17,
Jiao et al. in view of Dash et al. in view of Meilicke et al. and further in view of Gu teaches the AI device of claim 16.
In addition, Gu further teaches wherein each of the plurality of scores is a binary value of 1 or 0 (Equation 2.4; Section 2.2.2, first paragraph: "a comprehensive solution set contains all the qualified rules under certain local criteria. We denote by φ(P′) an indicator function that returns true if for any p ∈ P′, p is considered satisfactory to a collection of pre-defined local criteria, e.g., the coverage of positive instances or prediction confidence" teaches that quality scores for the rules are determined based on an indicator function (e.g. outputs binary score of 0 or 1) to determine if each rule is a qualified rule or not).
Jiao et al., Dash et al., Meilicke et al., and Gu are analogous to the claimed invention because they are directed towards knowledge graph link prediction.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate wherein each of the plurality of scores is a binary value of 1 or 0 as taught by Gu to the disclosed invention of Jiao et al. in view of Dash et al. and further in view of Meilicke et al.
One of ordinary skill in the art would have been motivated to make this modification because "a comprehensive solution set provides better predictive performance and interpretability than a concise set at the expense of model size and management overhead. The improvements in performance are mainly attributed to the fact that the rules in a comprehensive set can provide a more complete knowledge about target concepts" (Gu Section 2.2.2, first paragraph).
Regarding Claim 18,
Jiao et al. in view of Dash et al. in view of Meilicke et al. and further in view of Gu teaches the AI device of claim 16.
In addition, Gu further teaches wherein the controller is further configured to: compare the one quality score to a third predetermined threshold (Section 2.2.2, first paragraph: "a comprehensive solution set contains all the qualified rules under certain local criteria. We denote by φ(P′) an indicator function that returns true if for any p ∈ P′, p is considered satisfactory to a collection of pre-defined local criteria, e.g., the coverage of positive instances or prediction confidence" teaches comparing the satisfactory rule set from the learning system (one quality score) to some criteria (third predetermined threshold)), and
in response to the one quality score being greater than or equal to the third predetermined threshold, save the trained link prediction model in the memory of the Al device for deployment or transmit the trained link prediction model to an external device for deployment (Section 2.2, last paragraph: "given a target predicate r, a background knowledge B and sets of positive and negative instances of r, denoted by I+ and I− respectively, a rule learning algorithm aims to produce a set of rules P. When rules in P are grounded over B, CP(I+) and CP(I−) are satisfactory to specific criteria. The learned model is thus the rule set P" teaches that when the rule set determined by the rule learning algorithm is satisfactory to specific criteria (e.g. the one quality score is greater than or equal to a third predetermined threshold), the learned model is then saved as the rule set (e.g. trained model is saved in memory of AI device)).
Jiao et al., Dash et al., Meilicke et al., and Gu are analogous to the claimed invention because they are directed towards knowledge graph link prediction.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate wherein the controller is further configured to: compare the one quality score to a third predetermined threshold, and in response to the one quality score being greater than or equal to the third predetermined threshold, save the trained link prediction model in the memory of the Al device for deployment or transmit the trained link prediction model to an external device for deployment as taught by Gu to the disclosed invention of Jiao et al. in view of Dash et al. and further in view of Meilicke et al.
One of ordinary skill in the art would have been motivated to make this modification to "significantly reduces the runtime on evaluating instantiated rules, discovers much more high-quality rules than existing works and performs competitively on knowledge graph completion task compared to existing methods" (Gu Section 6, first paragraph).
Claims 9 and 19 are rejected under 35 U.S.C. 103 as being unpatentable over Jiao et al. (US 2022/0067030 A1) in view of Dash et al. (US 2024/0135205 A1) in view of Meilicke et al. ("Anytime Bottom-Up Rule Learning for Knowledge Graph Completion") and further in view Shang et al. (US 2020/0074301 A1).
Regarding Claim 9,
Jiao et al. in view of Dash et al. and further in view of Meilicke et al. teaches the method of claim 1.
Jiao et al. in view of Dash et al. and further in view of Meilicke et al. does not appear to explicitly teach further comprising: generating link prediction metrics based on the trained link prediction model, the link prediction metrics including at least one of mean rank (MR), mean reciprocal rank (MRR), and Hit@K; and determining whether to deploy the trained link prediction model based on the link prediction metrics.
However, Shang et al. teaches further comprising: generating link prediction metrics based on the trained link prediction model, the link prediction metrics including at least one of mean rank (MR), mean reciprocal rank (MRR), and Hit@K (Table 3; [0124]-[0125]: "Evaluation protocol Our experiments use the proportion of correct entities ranked in top 1,3 and 10 (Hits@1, Hits@3, Hits@10) and the mean reciprocal rank (MRR) as the metrics. In addition, since some corrupted triples may exist in the knowledge graphs, we use the filtered setting, i.e., we filter out all valid triples before ranking ... Link Prediction Our results on the standard FB15k-237, WN18RR and FB15k-237-Attr are shown in Table 3. Table 3 reports Hits@10, Hits@3, Hits@1 and MRR results of four different baseline models and two our models on three knowledge graphs datasets" teaches that Hit@k and mean reciprocal rank (MRR) metrics are generated for evaluating the trained model for link prediction); and
determining whether to deploy the trained link prediction model based on the link prediction metrics (Table 3; Fig. 10A; Fig. 10B; [0127]-[0129]: "Second, the structure information is added into our SACN model. In Table 3, SACN also get the best performances in the test dataset comparing all baseline methods. In FB15k-237, comparing ConvE, our SACN model improves Hits@10 value by a margin of 10.2%, Hits@3 value by a margin of 11.4%, Hits@1 value by a margin of 8.3% and MRR value by a margin of 9.4% for the test. In WN18RR dataset, comparing ConvE, our SACN model improves Hits@10 value by a margin of 12.5%, Hits@3 value by a margin of 11.6%, Hits@1 value by a margin of 10.3% and MRR value by a margin of 2.2% for the test ... Third, we add node attributes into our SACN model, i.e., we use the FB15k-237-Attr to train our model. The performance is improved again. Our model using attributes improves upon ConvE's Hits@10 by a margin of 12.2% , Hits@3 by a margin of 14.3%, Hits@1 by a margin of 12.5% and MRR by a margin of 12.5% ... Convergence Analysis FIG. 10A and FIG. 10B show the convergence of “Conv-TransE”, “SACN” and “SACN+Attr” models. We can see the SACN (red line) is always better than Conv-TransE (yellow line) after several epochs ... When using FB15k-237-Attr dataset, the performance of “SACN+Attr” is better than “SACN” model" teaches that the performance metrics (e.g. Hits@k and MRR) are used to determine the best model performance for deployment for link prediction (e.g. determining the model with the best performance metrics for deployment)).
Jiao et al., Dash et al., Meilicke et al., and Shang et al. are analogous to the claimed invention because they are directed towards knowledge graph link prediction.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate further comprising: generating link prediction metrics based on the trained link prediction model, the link prediction metrics including at least one of mean rank (MR), mean reciprocal rank (MRR), and Hit@K; and determining whether to deploy the trained link prediction model based on the link prediction metrics as taught by Shang et al. to the disclosed invention of Jiao et al. in view of Dash et al. and further in view of Meilicke et al.
One of ordinary skill in the art would have been motivated to make this modification to "compare our ... model with the four baseline models … [and] conclude that [our model] keeps the transitional characteristic between entities and relations and achieve better performance" (Shang et al. [0126]).
Regarding Claim 19,
Jiao et al. in view of Dash et al. and further in view of Meilicke et al. teaches the AI device of claim 11.
Jiao et al. in view of Dash et al. and further in view of Meilicke et al. does not appear to explicitly teach wherein the controller is further configured to: generate link prediction metrics based on the trained link prediction model, the link prediction metrics including at least one of mean rank (MR), mean reciprocal rank (MRR), and Hit@K, and determine whether to deploy the trained link prediction model based on the link prediction metrics.
However, Shang et al. teaches wherein the controller is further configured to: generate link prediction metrics based on the trained link prediction model, the link prediction metrics including at least one of mean rank (MR), mean reciprocal rank (MRR), and Hit@K (Table 3; [0124]-[0125]: "Evaluation protocol Our experiments use the proportion of correct entities ranked in top 1,3 and 10 (Hits@1, Hits@3, Hits@10) and the mean reciprocal rank (MRR) as the metrics. In addition, since some corrupted triples may exist in the knowledge graphs, we use the filtered setting, i.e., we filter out all valid triples before ranking ... Link Prediction Our results on the standard FB15k-237, WN18RR and FB15k-237-Attr are shown in Table 3. Table 3 reports Hits@10, Hits@3, Hits@1 and MRR results of four different baseline models and two our models on three knowledge graphs datasets" teaches that Hit@k and mean reciprocal rank (MRR) metrics are generated for evaluating the trained model for link prediction), and
determine whether to deploy the trained link prediction model based on the link prediction metrics (Table 3; Fig. 10A; Fig. 10B; [0127]-[0129]: "Second, the structure information is added into our SACN model. In Table 3, SACN also get the best performances in the test dataset comparing all baseline methods. In FB15k-237, comparing ConvE, our SACN model improves Hits@10 value by a margin of 10.2%, Hits@3 value by a margin of 11.4%, Hits@1 value by a margin of 8.3% and MRR value by a margin of 9.4% for the test. In WN18RR dataset, comparing ConvE, our SACN model improves Hits@10 value by a margin of 12.5%, Hits@3 value by a margin of 11.6%, Hits@1 value by a margin of 10.3% and MRR value by a margin of 2.2% for the test ... Third, we add node attributes into our SACN model, i.e., we use the FB15k-237-Attr to train our model. The performance is improved again. Our model using attributes improves upon ConvE's Hits@10 by a margin of 12.2% , Hits@3 by a margin of 14.3%, Hits@1 by a margin of 12.5% and MRR by a margin of 12.5% ... Convergence Analysis FIG. 10A and FIG. 10B show the convergence of “Conv-TransE”, “SACN” and “SACN+Attr” models. We can see the SACN (red line) is always better than Conv-TransE (yellow line) after several epochs ... When using FB15k-237-Attr dataset, the performance of “SACN+Attr” is better than “SACN” model" teaches that the performance metrics (e.g. Hits@k and MRR) are used to determine the best model performance for deployment for link prediction (e.g. determining the model with the best performance metrics for deployment)).
Jiao et al., Dash et al., Meilicke et al., and Shang et al. are analogous to the claimed invention because they are directed towards knowledge graph link prediction.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate wherein the controller is further configured to: generate link prediction metrics based on the trained link prediction model, the link prediction metrics including at least one of mean rank (MR), mean reciprocal rank (MRR), and Hit@K, and determine whether to deploy the trained link prediction model based on the link prediction metrics as taught by Shang et al. to the disclosed invention of Jiao et al. in view of Dash et al. and further in view of Meilicke et al.
One of ordinary skill in the art would have been motivated to make this modification to "compare our ... model with the four baseline models … [and] conclude that [our model] keeps the transitional characteristic between entities and relations and achieve better performance" (Shang et al. [0126]).
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to BRIAN J HALES whose telephone number is (571)272-0878. The examiner can normally be reached M-F 9:00am - 5:00pm.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Kamran Afshar can be reached at (571) 272-7796. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/BRIAN J HALES/Examiner, Art Unit 2125
/KAMRAN AFSHAR/Supervisory Patent Examiner, Art Unit 2125
Prosecution Insights