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 .
In response to Applicant’s claims filed on September 09, 2025, claims 1, 3-13 and 15-16 and 18-20 are now pending for examination in the application.
Response to Arguments
“The 112 rejection under 35 USC 112 set forth in the 09/09/2025 office action is hereby withdrawn.”
This office action is in response to amendment filed 09/09/2025. In this action claim(s) 1, 3-8, 16 and 18-20 is/are rejected under 35 U.S.C. 103 as being unpatentable over Mihindukulasooriya et al. (US Pub. No. 20220129770) in view of Manda et al. (US Pub. No. 20230153641). The Mihindukulasooriya et al. reference has been added to address the amendment of identifying, by a processing device, a query to retrieve data store in a data store comprising a data unit.
In this action claim(s) 9-13 and 15 is/are rejected under 35 U.S.C. 103 as being unpatentable over Miller et al. (US Pub. No. 20220253871) and Manda et al. (US Pub. No. 20230153641) in further view Mihindukulasooriya et al. (US Pub. No. 20220129770). The Mihindukulasooriya et al. reference has been added to address the amendment of receiving a query to retrieve data stored in a data store, the query comprising one or more query tokens.
Applicant’s arguments:
In regards to claim 1 on Page(s) 10, applicant argues “Claims 1, 3-13, 15-16, and 18—20 are not directed to an abstract idea because they recity a specific tehcnical solution involving multiple specialized machine learning models working in combination to solve the technical problem of standardizing and retrieving, from data stores, data that uses a variety of terms to describe similar content. The claimed solution includes generating and aggregating different types of embedding vectors that capture historical and lexical association in a way that cannot be performed mentally. The human mind is not capable of performing operations with embeddings.”
Examiner’s Reply:
Applicant argues that the claims comprises statutory subject matter. Examiner respectfully disagrees. The examiner notes that the computer as recited in the claims are being used for standardizing data for querying (the computer is being used as a generic tool). Therefore, the abstract idea recited in the claims is generally linking it to a computer environment, and does not integrate the abstract idea into a practical application. Generating and modifying Modeling data for data standardization for querying does not improve the functioning of a computing system. Receiving and retrieving data related to standardization is insignificant extra solution activity.
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, 3-13, 15-16, 18-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to non-patentable subject matter. The claims are directed to an abstract idea without significantly more.
Claim 1, 3-13, 15-16, 18-20 is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more. The judicial exception is not integrated into a practical application. The claims do not include additional elements that are sufficient to amount to significantly more than judicial exception. The eligibility analysis in support of these findings is provided below, on Claim Rejections - 35 USC 101 accordance with the "2019 Revised Patent Subject Matter Eligibility Guidance" (published on 1/7/2019 in Fed, Register, Vol. 84, No. 4 at pgs. 50-57, hereinafter referred to as the "2019 PEG").
Step 1. in accordance with Step 1 of the eligibility inquiry (as explained in MPEP 2106), it is first noted the claim method (claims 1, 3-8), a method (claim 9-13), a system (claim 16 and 18-20), are directed to one of the eligible categories of subject matter and therefore satisfies Step 1.
Step 2A. In accordance with Step 2A, prong one of the 2019 PEG, it is noted that the independent claims recite an abstract idea falling within the Mathematical Concepts & Mental Processes enumerated groupings of abstract ideas set forth in the 2019 PEG. Examiner is of the position that independent claims 1, 9, and 16 are directed towards the Mathematical Concepts & Mental Process Grouping of Abstract Ideas.
Independent claim(s) 1, 9, and 16 recites the following limitations directed towards a Mathematical Concepts & Mental Processes:
identifying, by a processing device, a query to retrieve data stored in a data store, the query comprising a data unit (This limitation, as drafted, is a process that, under its broadest reasonable interpretation, covers performance of the limitation in the mind but for the recitation of generic computer components. Nothing in the claim element precludes the step from practically being performed in the mind. For example, the claim encompasses identifying a query. Thus, the claim recites a mental process);
representing, by the processing device, the data unit via one or more tokens (This limitation, as drafted, is a process that, under its broadest reasonable interpretation, covers performance of the limitation in the mind but for the recitation of generic computer components. Nothing in the claim element precludes the step from practically being performed in the mind. For example, the claim encompasses representing with tokens. Thus, the claim recites a mental process);
processing, by the processing device, the one or more tokens using a plurality of machine learning models (MLMs) to identify affinity of the data unit to one or more clusters of a plurality of clusters, wherein each of the plurality of clusters is associated with one or more anchor tokens (This limitation, as drafted, is a process that, under its broadest reasonable interpretation, covers performance of the limitation in the mind but for the recitation of generic computer components. Nothing in the claim element precludes the step from practically being performed in the mind. For example, the claim encompasses identifying clusters. Thus, the claim recites a mental process), wherein processing the one or more tokens comprises:
processing the one or more tokens of the data unit using a statistical MLM of the plurality of MLMs to generate a first embedding vector comprising a first plurality of components, wherein each component of the first plurality of components characterizes a historical association of the one or more tokens of the data unit with the one or more anchor tokens of a respective cluster of the plurality of clusters (This limitation, as drafted, is a process that, under its broadest reasonable interpretation, covers performance of the limitation in the mind but for the recitation of generic computer components. Nothing in the claim element precludes the step from practically being performed in the mind. For example, the claim encompasses generating a vector. Thus, the claim recites a mental process); and
processing the one or more tokens of the data unit using a natural language processing MLM of the plurality of MLMs to generate a second embedding vector comprising a second plurality of components, wherein each component of the second plurality of components characterizes a lexical association of the one or more tokens of the data unit with the one or more anchor tokens of the respective cluster of the plurality of clusters (This limitation, as drafted, is a process that, under its broadest reasonable interpretation, covers performance of the limitation in the mind but for the recitation of generic computer components. Nothing in the claim element precludes the step from practically being performed in the mind. For example, the claim encompasses generating a vector. Thus, the claim recites a mental process); and
identifying affinity of the data unit to the one or more clusters of the plurality of clusters using a third embedding vector, aggregated from the first embedding vector and the second embedding vector (This limitation, as drafted, is a process that, under its broadest reasonable interpretation, covers performance of the limitation in the mind but for the recitation of generic computer components. Nothing in the claim element precludes the step from practically being performed in the mind. For example, the claim encompasses identifying a cluster. Thus, the claim recites a mental process).
Step 2A. In accordance with Step 2A, prong two of the 2019 PEG, the judicial exception is not integrated into a practical application because of the recitation in claim(s) 1, 9, and 16:
processing, by the processing device, the one or more tokens using a plurality of machine learning models (MLMs) to identify one or more clusters of a plurality of clusters, wherein each of the one or more identified clusters is associated with at least one token of the one or more tokens ((machine learning models (i.e., merely automate the claimed steps and are no more than mere instructions to apply the exception using generic computer components), wherein processing the one or more tokens comprises:
processing the one or more tokens using a first MLM of the plurality of MLMs to generate a first embedding vector comprising a first plurality of components, wherein an individual component of the first plurality of components characterizes a likelihood of a historical association of the one or more tokens with a respective cluster of the plurality of clusters (machine learning models (i.e., merely automate the claimed steps and are no more than mere instructions to apply the exception using generic computer components); and
processing the one or more tokens using a second MLM of the plurality of MLMs to generate a second embedding vector comprising a second plurality of components, wherein an individual component of the second plurality of components characterizes a likelihood of a lexical association of the one or more tokens with a respective cluster of the plurality of clusters ((machine learning models (i.e., merely automate the claimed steps and are no more than mere instructions to apply the exception using generic computer components); and
retrieving, using the one or more clusters, the data stored in the data store in association with the one or more clusters (recites insignificant extra solution activity that amounts to retrieving query data).
The claim as a whole merely describes how to generally “apply” the exception in a
computer environment. Even when viewed in combination, these additional elements do not integrate the abstract idea into a practical application because they do not impose any
meaningful limits on practicing the abstract idea. The claim is directed to the abstract idea.
Step 2B. Similar to the analysis under 2A Prong Two, the claim(s) does/do not include
additional elements that are sufficient to amount to significantly more than the judicial
exception. Because the additional elements of the independent claims amount to insignificant
extra solution activity and/or mere instructions, the additional elements do not add significantly more to the judicial exception such that the independent claims as a whole would be patent eligible.
Therefore, independent claims 1, 15, and 16 are rejected under 35 U.S.C. 101.
With respect to claim(s) 3 and 18:
Step 2A, prong one of the 2019 PEG:
wherein the historical association of the one or more tokens of the data unit with the one or more anchor tokens of the respective cluster of the plurality of clusters characterizes a number of times the one or more tokens of the data unit have been encountered in training data units together with the one or more anchor tokens of the respective given cluster (The limitation recites a mental process of observation and/or evaluation capable of being performed by the human mind by generating a vector).
Step 2A Prong Two Analysis:
This judicial exception is not integrated into a practical application because there are no
additional elements to provide practical application.
Step 2B Analysis:
The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. The claim is not patent eligible.
With respect to claim(s) 4:
Step 2A, prong one of the 2019 PEG:
wherein the training data units comprise data units previously processed by the statistical MLM (The limitation recites a mental process of observation and/or evaluation capable of being performed by the human mind by generating a vector).
Step 2A Prong Two Analysis:
This judicial exception is not integrated into a practical application because there are no
additional elements to provide practical application.
Step 2B Analysis:
The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. The claim is not patent eligible.
With respect to claim(s) 5:
Step 2A, prong one of the 2019 PEG:
updating one or more parameters of the statistical MLM in view of the one or more tokens of the data unit and at least a subset of the plurality of clusters (The limitation recites a mental process of observation and/or evaluation capable of being performed by the human mind by updating a parameter).
Step 2A Prong Two Analysis:
This judicial exception is not integrated into a practical application because there are no
additional elements to provide practical application.
Step 2B Analysis:
The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. The claim is not patent eligible.
With respect to claim(s) 6 and 19:
Step 2A, prong one of the 2019 PEG:
wherein each cluster of at least a subset of the plurality of clusters is associated with different lexical units having a substantially same semantic meaning (The limitation recites a mental process of observation and/or evaluation capable of being performed by the human mind by identifying a cluster).
Step 2A Prong Two Analysis:
This judicial exception is not integrated into a practical application because there are no
additional elements to provide practical application.
Step 2B Analysis:
The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. The claim is not patent eligible.
With respect to claim(s) 7:
Step 2A, prong one of the 2019 PEG:
a plurality of training data units, and
ground truth labels generated by application of the statistical MLM to the plurality of training data units (The limitation recites a mental process of observation and/or evaluation capable of being performed by the human mind by identifying a cluster).
Step 2A Prong Two Analysis:
This judicial exception is not integrated into a practical application because there are no
additional elements to provide practical application.
Step 2B Analysis:
The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. The claim is not patent eligible.
With respect to claim(s) 8 and 20:
Step 2A, prong one of the 2019 PEG:
correcting spelling of one or more words in the data unit (The limitation recites a mental process of observation and/or evaluation capable of being performed by the human mind by coreccting spelling),
translating one or more foreign-language words in the data unit (The limitation recites a mental process of observation and/or evaluation capable of being performed by the human mind by translating words),
replacing one or more acronyms in the data unit (The limitation recites a mental process of observation and/or evaluation capable of being performed by the human mind by coreccting spelling),
expanding one or more abbreviations in the data unit (The limitation recites a mental process of observation and/or evaluation capable of being performed by the human mind by expanding abbreviations), or
adding one or more spaces between two or more words in the data unit (The limitation recites a mental process of observation and/or evaluation capable of being performed by the human mind by spacing words).
Step 2A Prong Two Analysis:
This judicial exception is not integrated into a practical application because there are no
additional elements to provide practical application.
Step 2B Analysis:
The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. The claim is not patent eligible.
With respect to claim(s) 10:
Step 2A, prong one of the 2019 PEG:
Examiner is of the position the dependent claim is directed toward additional elements.
Step 2A Prong Two Analysis:
causing the first MLM to access a second cluster score of each token of the one or more tokens (recites insignificant extra solution activity that amounts to mere accessing cluster data); and
causing the first MLM to modify the second cluster score of each token of the one or more tokens in view of a second number of occurrences in the training data unit of the one or more anchor tokens assigned to a second cluster of the plurality of clusters (recites insignificant extra solution activity that amounts to mere accessing cluster data).
Step 2B Analysis:
The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. The specification does not provide any indication that the accessing cluster, and its contents, was done by anything other than a generic, off-the-shelf computer component. Additionally, the Symantec, TLI, and OIP Techs court decisions cited in MPEP 2106.05(d)(II) indicate that mere collection or receipt of data over a network is a well- understood, routine, and conventional function when it is claimed in a merely generic manner (as it is here). Accordingly, a conclusion that this limitation is a well-understood, routine, and conventional activity is supported. The claim is not patent eligible.
With respect to claim(s) 11:
Step 2A, prong one of the 2019 PEG:
wherein each cluster of at least a subset of the plurality of clusters is associated with different lexical units having a substantially same semantic meaning (The limitation recites a mental process of observation and/or evaluation capable of being performed by the identifying cluster-related data).
Step 2A Prong Two Analysis:
This judicial exception is not integrated into a practical application because there are no
additional elements to provide practical application.
Step 2B Analysis:
The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. The claim is not patent eligible.
With respect to claim(s) 12:
Step 2A, prong one of the 2019 PEG:
wherein, prior to processing training the first MLM using the training data unit, the first cluster score characterizes a number of times a respective token of the one or more tokens has been encountered in previous training data units together with the one or more anchor tokens assigned to the first cluster (The limitation recites a mental process of observation and/or evaluation capable of being performed by the scoring cluster-related data).
Step 2A Prong Two Analysis:
This judicial exception is not integrated into a practical application because there are no
additional elements to provide practical application.
Step 2B Analysis:
The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. The claim is not patent eligible.
With respect to claim(s) 13:
Step 2A, prong one of the 2019 PEG:
wherein a higher first cluster score indicates a higher probability of occurrence, in a same data unit, of a respective token of the one or more tokens with the one or more anchor tokens assigned to the first cluster (The limitation recites a mental process of observation and/or evaluation capable of being performed by the scoring cluster-related data).
Step 2A Prong Two Analysis:
This judicial exception is not integrated into a practical application because there are no
additional elements to provide practical application.
Step 2B Analysis:
The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. The claim is not patent eligible.
With respect to claim(s) 15:
Step 2A, prong one of the 2019 PEG:
training a second MLM of the one or more MLMs using a plurality of training data units and ground truth labels generated by application of the first MLM to the plurality of training data units, wherein the second MLM is a pre-trained embeddings language model (The limitation recites a mental process of observation and/or evaluation capable of being performed by the identifying cluster data);
causing the second MLM to generate a second embedding vector comprising a second plurality of components, wherein an individual component of the second plurality of components characterizes a likelihood of a lexical association of the one or more query tokens with the one or more anchor tokens of a respective cluster of the plurality of clusters (The limitation recites a mental process of observation and/or evaluation capable of being performed by the generating a vector).
Step 2A Prong Two Analysis:
wherein to obtain the plurality of probabilities, the machine learning classifier is further to process the second embedding vector (recites insignificant extra solution activity that amounts to obtaining probability data).
Step 2B Analysis:
The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. The claim is not patent eligible.
Claim Rejections - 35 USC § 103
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.
Claim(s) 1, 3-8, 16 and 18-20 is/are rejected under 35 U.S.C. 103 as being unpatentable over
Mihindukulasooriya et al. (US Pub. No. 20220129770) in view of Manda et al. (US Pub. No. 20230153641).
With respect to claim 1, Mihindukulasooriya et al. teaches a method comprising:
identifying, by a processing device, a query to retrieve data store in a data store comprising a data unit (Paragraph 20 discloses identifying, by the processor, a natural language query; translating, by the processor, the natural language query into an intermediate representation; converting, by the processor, the intermediate representation into one or more query triples and Paragraph 55 discloses the knowledge base may be stored within one or more data stores (e.g., databases, etc.));
representing, by the processing device, the data unit via one or more tokens (Paragraph 58 discloses the natural language sentence includes at least a predetermined number of terms (e.g., tokens, etc.) and at least one verb);
processing, by the processing device, the one or more tokens using a plurality of machine learning models (MLMs) to identify affinity of the data unit to one or more clusters of a plurality of clusters , wherein each of the plurality of clusters is associated with one or more anchor tokens (Paragraph 62 discloses the machine learning environment may be trained to assign a high similarity score to a relationship between each of the linked triples), wherein processing the one or more tokens comprises:
processing the one or more tokens of the data unit using a statistical MLM of the plurality of MLMs to generate a first embedding vector comprising a first plurality of components, wherein each component of the first plurality of components characterizes a historical association of the one or more tokens of the data unit with the one or more anchor tokens of a respective cluster of the plurality of clusters (Paragraph 122 discloses a neural model is trained for relation linking by exploiting the distant supervision dataset. The neural model produces dense embedding vectors for input questions, which can learn to project the same relation type's different surface forms close in the latent space);
processing the one or more tokens of the data unit using a natural language processing MLM of the plurality of MLMs to generate a second embedding vector comprising a second plurality of components, wherein an individual component of the second plurality of components characterizes a lexical association of the one or more tokens of the data unit with the one or more anchor tokens of the respective cluster of the plurality of clusters (Paragraph 125 discloses vectors of the final-layer hidden states of the start entity markers of subject and object entities are concatenated and fed into a fully connected layer to get the final embedding vector for the relation instance x. Singh et al. does not disclose processing the one or more tokens using a first MLM of the plurality of MLMs to generate a first embedding vector comprising a first plurality of components, wherein an individual component of the first plurality of components characterizes a historical association of the one or more tokens with a respective cluster of the plurality of clusters.
However, Manda et al. teaches processing the one or more tokens of the data unit using a statistical MLM of the plurality of MLMs to generate a first embedding vector comprising a first plurality of components, wherein each component of the first plurality of components characterizes a likelihood of a historical association of the one or more tokens with a respective cluster of the plurality of clusters (Paragraph 48 discloses the extraction engine 150 can generate performance metrics and/or other historical feedback for a particular machine learning model and store this data relationally to the inputs and/or outputs of the machine learning model);
Identifying affinity of the data unit to the one or more clusters of the plurality of clusters using a third embedding vector, aggregated from the first embedding vector and the second embedding vector (Paragraph 56 discloses extractive summarization can include the ability to extract key information from public news articles, to produce insights such as trends and news spotlights, ability to classify documents by their key contents, ability to distill important information from long documents to empower solutions such as search, question and answer formats and decision support, ability to cluster documents by their relevant content); and
retrieving, using the one or more clusters, the data stored in the data store in association with the one or more clusters (Paragraph 56 discloses extractive summarization can include the ability to extract key information from public news articles, to produce insights such as trends and news spotlights, ability to classify documents by their key contents, ability to distill important information from long documents to empower solutions such as search, question and answer formats and decision support, ability to cluster documents by their relevant content).
Therefore, it would have been obvious at the time the invention was made to a person having ordinary skill in the art to modify over Mihindukulasooriya et al. with Manda et al. to processing the one or more tokens using a first MLM of the plurality of MLMs to generate a first embedding vector comprising a first plurality of components, wherein an individual component of the first plurality of components characterizes a likelihood of a historical association of the one or more tokens with a respective cluster of the plurality of clusters. This would have facilitated data standardization.
The Mihindukulasooriya et al. reference as modified by Manda et al. teaches all the limitations of claim 2. Regarding claim 3, Manda et al. discloses the method of claim 2, wherein the historical association of the one or more tokens with the respective cluster of the plurality of clusters characterizes a number of times the one or more tokens have been encountered in training data units together with the one or more anchor tokens associated with the respective given cluster (Paragraph 48 discloses the extraction engine 150 can generate performance metrics and/or other historical feedback for a particular machine learning model and store this data relationally to the inputs and/or outputs of the machine learning model). The motivation to combine statement previously provided in the rejection of dependent claim 2 provided above, combining the Mihindukulasooriya et al. reference and the Manda et al. reference is applicable to dependent claim 3.
The Mihindukulasooriya et al. reference as modified by Manda et al. teaches all the limitations of claim 3. Regarding claim 4, Mihindukulasooriya et al. discloses the method of claim 3, wherein the training data units comprise data units previously processed by the first MLM (Paragraph 62 discloses the machine learning environment may be trained to assign a high similarity score to a relationship between each of the linked triples. This may improve an accuracy of the machine learning environment. In another embodiment, the machine learning environment may include one or more deep learning implementations, one or more neural networks, etc).
The Mihindukulasooriya et al. reference as modified by Manda et al. teaches all the limitations of claim 3. Regarding claim 5, Mihindukulasooriya et al. discloses the method of claim 3, further comprising:
updating one or more parameters of the first MLM in view of the one or more tokens and the one or more identified clusters (Paragraph 58-59 discloses a given knowledge base triple, a natural language sentence may be selected where the subject and object of the knowledge base triple co-occur within the natural language sentence (and do not overlap as a single word), and the natural language sentence includes at least a predetermined number of terms (e.g., tokens, etc.) and at least one verb. Furthermore, method 500 may proceed with operation 506, where a machine learning environment is trained utilizing the plurality of knowledge base triples and their corresponding natural language sentences).
The Mihindukulasooriya et al. reference as modified by Manda et al. teaches all the limitations of claim 1. Regarding claim 6, Mihindukulasooriya et al. discloses the method of claim 1, wherein each cluster of at least a subset of the plurality of clusters is associated with different lexical units having a substantially same semantic meaning (Paragraph 65 discloses the ambiguity of natural language and lack of training data. To overcome these challenges, SLING is provided, which includes a relation linking framework which leverages semantic parsing using Abstract Meaning Representation (AMR) and distant supervision).
The Mihindukulasooriya et al. reference as modified by Manda et al. teaches all the limitations of claim 1. Regarding claim 7, Mihindukulasooriya et al. discloses the method of claim 1, wherein the natural language processing MLM is trained using:
a plurality of training data units, and ground truth labels generated by application of the first MLM to the plurality of training data units (Paragraph 145 discloses relation linking between a knowledge base (KB) and a semantic parse (e.g., one or more abstract meaning representation (AMR) graphs created using semantic role labeling (SRL)) using a parallel corpus with distant supervision).
The Mihindukulasooriya et al. reference as modified by Manda et al. teaches all the limitations of claim 1. Regarding claim 8, Manda et al. discloses the method of claim 1, wherein representing the data unit via one or more tokens comprises performing at least one of:
correcting spelling of one or more words in the data unit, translating one or more foreign-language words in the data unit, replacing one or more acronyms in the data unit,
expanding one or more abbreviations in the data unit, or adding one or more spaces between two or more words in the data unit (Paragraph 64 discloses In some embodiments, post-processing operations at 406 can include operations to pass entities through a spell checker module to fix spelling issues by, for example, referencing the relevant ontology. In some embodiments, post-processing operations at 406 can include operations to map spell-corrected medication entities to fetch corresponding RxCUI codes). The motivation to combine statement previously provided in the rejection of dependent claim 1 provided above, combining the Mihindukulasooriya et al. reference and the Manda et al. reference is applicable to dependent claim 8.
With respect to claim 16, Mihindukulasooriya et al. teaches a system comprising:
a memory (See Fig. 2); and
a processing device coupled to the memory (See Fig. 2), the processing device to perform operations comprising:
identifying, by a processing device, a query to retrieve data store in a data store comprising a data unit (Paragraph 20 discloses identifying, by the processor, a natural language query; translating, by the processor, the natural language query into an intermediate representation; converting, by the processor, the intermediate representation into one or more query triples and Paragraph 55 discloses the knowledge base may be stored within one or more data stores (e.g., databases, etc.));
representing, by the processing device, the data unit via one or more tokens (Paragraph 58 discloses the natural language sentence includes at least a predetermined number of terms (e.g., tokens, etc.) and at least one verb);
processing, by the processing device, the one or more tokens using a plurality of machine learning models (MLMs) to identify affinity of the data unit to one or more clusters of a plurality of clusters , wherein each of the plurality of clusters is associated with one or more anchor tokens (Paragraph 62 discloses the machine learning environment may be trained to assign a high similarity score to a relationship between each of the linked triples), wherein processing the one or more tokens comprises:
processing the one or more tokens of the data unit using a statistical MLM of the plurality of MLMs to generate a first embedding vector comprising a first plurality of components, wherein each component of the first plurality of components characterizes a historical association of the one or more tokens of the data unit with the one or more anchor tokens of a respective cluster of the plurality of clusters (Paragraph 122 discloses a neural model is trained for relation linking by exploiting the distant supervision dataset. The neural model produces dense embedding vectors for input questions, which can learn to project the same relation type's different surface forms close in the latent space);
processing the one or more tokens of the data unit using a natural language processing MLM of the plurality of MLMs to generate a second embedding vector comprising a second plurality of components, wherein an individual component of the second plurality of components characterizes a lexical association of the one or more tokens of the data unit with the one or more anchor tokens of the respective cluster of the plurality of clusters (Paragraph 125 discloses vectors of the final-layer hidden states of the start entity markers of subject and object entities are concatenated and fed into a fully connected layer to get the final embedding vector for the relation instance x. Singh et al. does not disclose processing the one or more tokens using a first MLM of the plurality of MLMs to generate a first embedding vector comprising a first plurality of components, wherein an individual component of the first plurality of components characterizes a historical association of the one or more tokens with a respective cluster of the plurality of clusters.
However, Manda et al. teaches processing the one or more tokens of the data unit using a statistical MLM of the plurality of MLMs to generate a first embedding vector comprising a first plurality of components, wherein each component of the first plurality of components characterizes a likelihood of a historical association of the one or more tokens with a respective cluster of the plurality of clusters (Paragraph 48 discloses the extraction engine 150 can generate performance metrics and/or other historical feedback for a particular machine learning model and store this data relationally to the inputs and/or outputs of the machine learning model);
Identifying affinity of the data unit to the one or more clusters of the plurality of clusters using a third embedding vector, aggregated from the first embedding vector and the second embedding vector (Paragraph 56 discloses extractive summarization can include the ability to extract key information from public news articles, to produce insights such as trends and news spotlights, ability to classify documents by their key contents, ability to distill important information from long documents to empower solutions such as search, question and answer formats and decision support, ability to cluster documents by their relevant content); and
retrieving, using the one or more clusters, the data stored in the data store in association with the one or more clusters (Paragraph 56 discloses extractive summarization can include the ability to extract key information from public news articles, to produce insights such as trends and news spotlights, ability to classify documents by their key contents, ability to distill important information from long documents to empower solutions such as search, question and answer formats and decision support, ability to cluster documents by their relevant content).
Therefore, it would have been obvious at the time the invention was made to a person having ordinary skill in the art to modify over Mihindukulasooriya et al. with Manda et al. to processing the one or more tokens using a first MLM of the plurality of MLMs to generate a first embedding vector comprising a first plurality of components, wherein an individual component of the first plurality of components characterizes a likelihood of a historical association of the one or more tokens with a respective cluster of the plurality of clusters. This would have facilitated data standardization.
With respect to claim 18, it is rejected on grounds corresponding to above rejected claim 3, because claim 18 is substantially equivalent to claim 3.
With respect to claim 19, it is rejected on grounds corresponding to above rejected claim 6, because claim 19 is substantially equivalent to claim 6.
With respect to claim 20, it is rejected on grounds corresponding to above rejected claim 8, because claim 20 is substantially equivalent to claim 8.
Claim(s) 9-13 and 15 is/are rejected under 35 U.S.C. 103 as being unpatentable over Miller et al. (US Pub. No. 20220253871) and Manda et al. (US Pub. No. 20230153641) in further view Mihindukulasooriya et al. (US Pub. No. 20220129770).
With respect to claim 9, Miller et al. teaches a method to train one or more machine learning models (MLMs), the method comprising:
initiating a plurality of clusters by assigning one or more anchor tokens to each cluster of the plurality of clusters (Paragraph 162 discloses ML models comprise method(s) often used in unsupervised or reinforced ML methods such as k-means (or variants thereof, such as K means++)/nearest neighbor analytical models, such as k-nearest neighbor analysis; other clustering methods (e.g., partitional clustering, mean shift clustering, density based clustering (e.g., DBSCAN methods), or hierarchical clustering (such as agglomerative clustering));
training a first MLM using a training data unit comprising one or more tokens (Paragraph 316 disclose MLMs can be dynamically updated over time through feeding of updated training data), wherein training the first MLM comprises:
causing the first MLM to access a first cluster score of each token of the one or more tokens, the first cluster score characterizing affinity of a respective token with the one or more anchor tokens of a first cluster of the plurality of clusters (Paragraph 162 discloses affinity mapping and Paragraph 339 discloses NLP processes (e.g., tokenization, delimiter identification, sentence/phrase identification, structure determination, etc.) (e.g., #7), and comparison with SN or corpus/rule(s) for determination of specialized terms). Miller et al. does not disclose causing the first MLM to modify the first cluster score of each token of the one or more tokens in view of a first number of occurrences, in the training data unit, of the one or more anchor tokens assigned to the first cluster of the plurality of clusters.
However, teaches Manda et al. causing the first MLM to modify the first cluster score of each token of the one or more tokens in view of a first number of occurrences, in the training data unit, of the one or more anchor tokens assigned to the first cluster of the plurality of clusters (Paragraph 66 discloses confidence scores, expected similarity scores, or similar metrics, which can improve over time as the model is trained on additional data and Paragraph 56 discloses cluster documents by their relevant content, and ability to highlight key sentences in documents);
causing the first MLM to generate, using the first cluster scores of each of the one or more query tokens, an embedding vector comprising a plurality of components, wherein each component of the plurality of components characterizes a likelihood of historical association of the one or more query tokens with the one or more anchor tokens of a respective cluster of the plurality of clusters (Paragraph 48 discloses the extraction engine 150 can generate performance metrics and/or other historical feedback for a particular machine learning model and store this data relationally to the inputs and/or outputs of the machine learning model).
Therefore, it would have been obvious at the time the invention was made to a person having ordinary skill in the art to modify over Miller et al. with Manda et al. to causing the first MLM to access a first cluster score of each token of the one or more tokens. This would have facilitated data standardization.
Miller et al. as modified by Manda et al. does not disclose processing, using a machine learning classifier, the embedding vector to obtain a plurality of probabilities characterizing association of the query to the plurality of clusters.
However, Mihindukulasooriya et al. teaches receiving a query to retrieve data stored in a data store, the query comprising one or more query tokens (Paragraph 20 discloses identifying, by the processor, a natural language query; translating, by the processor, the natural language query into an intermediate representation; converting, by the processor, the intermediate representation into one or more query triples and Paragraph 55 discloses the knowledge base may be stored within one or more data stores (e.g., databases, etc.));
processing, using a machine learning classifier, the embedding vector to obtain a plurality of probabilities characterizing association of the query to the plurality of clusters (See Paragraph 122 discloses neural model produces dense embedding vectors for input questions and Paragraph 124 discloses a neural network M is trained on D with the purpose to predict the correct relation type r.sub.k given the instance x.sub.k by minimizing the cross-entropy loss regarding the conditional probability p.sub.M(⋅|x.sub.k) modeled by M, with respect to the true relation t.sub.k); and
retrieving, using the plurality of probabilities, the data stored in the data store (See Paragraph 152 discloses a document collection (DC), a question in a natural language, a placeholder for the subject, and a placeholder for the object. The method provides as output a list of relation types sorted by probability scores).
Therefore, it would have been obvious at the time the invention was made to a person having ordinary skill in the art to modify over Miller et al. and Manda et al. with Mihindukulasooriya et al. to process, using a machine learning classifier, the embedding vector to obtain a plurality of probabilities characterizing association of the query to the plurality of clusters. This would have facilitated data standardization.
The Miller et al. reference as modified by Manda et al. and Mihindukulasooriya et al. teaches all the limitations of claim 9. Regarding claim 10, Manda et al. discloses the method of claim 9, wherein processing the training the first MLM data unit further comprises:
causing the first MLM to access a second cluster score of each token of the one or more tokens (Paragraph 66 discloses confidence scores, expected similarity scores, or similar metrics, which can improve over time as the model is trained on additional data and Paragraph 56 discloses cluster documents by their relevant content, and ability to highlight key sentences in documents); and
causing the first MLM to modify the second cluster score of each token of the one or more tokens in view of a second number of occurrences in the training data unit of the one or more anchor tokens assigned to a second cluster of the plurality of clusters (Paragraph 66 discloses confidence scores, expected similarity scores, or similar metrics, which can improve over time as the model is trained on additional data and Paragraph 56 discloses cluster documents by their relevant content, and ability to highlight key sentences in documents). The motivation to combine statement previously provided in the rejection of dependent claim 9 provided above, combining the Miller et al. reference and the Manda et al. reference is applicable to dependent claim 10.
The Miller et al. reference as modified by Manda et al. and Mihindukulasooriya et al. teaches all the limitations of claim 9. Regarding claim 11, Miller et al. discloses the method of claim 9, wherein each cluster of at least a subset of the plurality of clusters is associated with different lexical units having a substantially same semantic meaning (Paragraph 60 discloses semantic term vectors and lexical term vectors).
The Miller et al. reference as modified by Manda et al. and Mihindukulasooriya et al. teaches all the limitations of claim 9. Regarding claim 12, Manda et al. discloses the method of claim 9, wherein, prior to processing training the first MLM using the training data unit, the first cluster score characterizes a number of times a respective token of the one or more tokens has been encountered in previous training data units together with the one or more anchor tokens assigned to the first cluster (Paragraph 48 discloses the extraction engine 150 can generate performance metrics and/or other historical feedback for a particular machine learning model and store this data relationally to the inputs and/or outputs of the machine learning model). The motivation to combine statement previously provided in the rejection of dependent claim 9 provided above, combining the Miller et al. reference and the Manda et al. reference is applicable to dependent claim 12.
The Miller et al. reference as modified by Manda et al. and Mihindukulasooriya et al. teaches all the limitations of claim 9. Regarding claim 13, Manda et al. discloses the method of claim 9, wherein a higher first cluster score indicates a higher probability of occurrence, in a same data unit, of a respective token of the one or more tokens with the one or more anchor tokens assigned to the first cluster (Paragraph 66 discloses confidence scores, expected similarity scores, or similar metrics, which can improve over time as the model is trained on additional data and Paragraph 56 discloses cluster documents by their relevant content, and ability to highlight key sentences in documents). The motivation to combine statement previously provided in the rejection of dependent claim 9 provided above, combining the Kanagovi et al. reference and the Manda et al. reference is applicable to dependent claim 13.
The Miller et al. reference as modified by Manda et al. and Mihindukulasooriya et al. teaches all the limitations of claim 9. Regarding claim 15, Miller et al. discloses the method of claim 9, further comprising:
training a second MLM of the one or more MLMs using a plurality of training data units and ground truth labels generated by application of the first MLM to the plurality of training data units, wherein the second MLM is a pre-trained embeddings language model (Paragraph 161 discloses development of MLM method(s) comprise the steps of applying Feature learning method(s), Feature engineering methods(s), or both to Function(s)/DS(s) to develop MLM method(s); applying supervised or semi-supervised learning/refinement of such MLM-implemented Function(s); applying reinforced learning, unsupervised learning to enhance such Function(s); and eventually allowing Function(s) or aspects of Function(s) to be governed by the trained model. In aspects, ML-implemented Functions (“MLIFs”) characterize/catalog DSF(s), Records, Customers, etc. In aspects, MLIFs identify patterns or relationships).
causing the second MLM to generate a second embedding vector comprising a second plurality of components, wherein each component of the second plurality of components characterizes a likelihood of a lexical association of the one or more query tokens with the one or more anchor tokens of a respective cluster of the plurality of clusters(Paragraph 161 discloses development of MLM method(s) comprise the steps of applying Feature learning method(s), Feature engineering methods(s), or both to Function(s)/DS(s) to develop MLM method(s); applying supervised or semi-supervised learning/refinement of such MLM-implemented Function(s); applying reinforced learning, unsupervised learning to enhance such Function(s); and eventually allowing Function(s) or aspects of Function(s) to be governed by the trained model. In aspects, ML-implemented Functions (“MLIFs”) characterize/catalog DSF(s), Records, Customers, etc. In aspects, MLIFs identify patterns or relationships); and
wherein to obtain the plurality of probabilities, the machine learning classifier is further to process the second embedding vector (Paragraph 161 discloses development of MLM method(s) comprise the steps of applying Feature learning method(s), Feature engineering methods(s), or both to Function(s)/DS(s) to develop MLM method(s); applying supervised or semi-supervised learning/refinement of such MLM-implemented Function(s); applying reinforced learning, unsupervised learning to enhance such Function(s); and eventually allowing Function(s) or aspects of Function(s) to be governed by the trained model. In aspects, ML-implemented Functions (“MLIFs”) characterize/catalog DSF(s), Records, Customers, etc. In aspects, MLIFs identify patterns or relationships).
Relevant Prior Art
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure. The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
US PG-PUB 20200097587 is directed to MACHINE LEARNING DETECTION OF DATABASE INJECTION ATTACKS: [0035] automatically detect query language statements that may include a SQL injection attack, or otherwise represent malicious code. In a first method, a machine learning technique is trained by comparing two versions of a query language statement. A version is created that uses processed user input, such as user input processed using parameter (or user input) escaping, and a version is created that is not processed, such as not being escaped. Tokens, and optionally relationships between tokens, produced by the two versions are compared. Comparison can include a token-by-token comparison, or calculating hash values of sets of tokens, with a SQL injection attempt implicated if different hash values are produced. If the versions differ, a SQL injection attempt can be presumed, and the query, or a portion thereof (e.g., the user input, the pattern of escaping, tokens included in the query and their order) can be used (as part of a machine learning model) to analyze test data (e.g., data to be classified using the trained classifier).
Conclusion
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to NICHOLAS E ALLEN whose telephone number is (571)270-3562. The examiner can normally be reached Monday through Thursday 830-630.
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, Boris Gorney can be reached at (571) 270-5626. 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.
/N.E.A/Examiner, Art Unit 2154
/BORIS GORNEY/Supervisory Patent Examiner, Art Unit 2154
Prosecution Insights