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 .
DETAILED ACTION
This action is in response to the application filed 02 March 2026. Claims 9 and 10 are canceled. Claims 1, 11, 12, 16, and 20 are amended. Claims 1-8 and 11-20 are pending and have been examined.
Continued Examination Under 37 CFR 1.114
A request for continued examination under 37 CFR 1.114, including the fee set forth in 37 CFR 1.17(e), was filed in this application after final rejection. Since this application is eligible for continued examination under 37 CFR 1.114, and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114. Applicant's submission filed on 02 March 2026 has been entered.
Response to Arguments
Applicant's arguments, see pages 11-14, filed 02 March 2026, with respect to the rejections of Claims 1-20 under 35 U.S.C. 101 have been fully considered but they are not persuasive.
APPLICANT'S ARGUMENT: Applicant argues (page 12, paragraphs 1-3) that "the claimed invention, specifies a process that is far from such a generic, human-performable activity. ¶ As recited above, the combination of features claimed defines a highly specific algorithmic process that is not practically feasible for a human to perform. The claims recite features that represent a complex computational task. ¶ This process involves sophisticated data handling, combinatorial problem-solving, and precise mathematical operations on a scale and at a speed beyond human capabilities. ... The intricate balancing of distinct cost types with differential weights, specifically designed to address both logistical efficiency and human-centric factors, is not a mere mental exercise but a complex optimization procedure intrinsically tied to computer implementation."
Applicant argues (page 14, paragraph 3) that "The claimed process is not merely an automation of a known mental process or the application of an abstract idea on a generic computer. Instead, it defines a specific algorithmic approach for optimizing a complex real-world problem."
EXAMINER'S RESPONSE: Examiner respectfully disagrees. As currently claimed, amended Claim 1 appears merely to recite a mental process that is implemented by means of computing technology without claiming an improvement to computing technology or the functioning of a computer. Applicant's purported technological improvement by means of a "highly specific algorithmic process," "combinatorial problem-solving," and "precise mathematical operations on a scale and at a speed beyond human capabilities" must also recite additional elements that integrate the mental process steps into a practical application or provide significantly more in order not to be directed to the recited mental processes.
Examiner further notes that amended Claim 1 does not appear to recite or reflect "distinct cost types with differential weights" or "logistical efficiency." Although the claims are interpreted in light of the specification, limitations from the specification are not read into the claims.
APPLICANT'S ARGUMENT: Applicant argues (page 13, paragraphs 1-3) that "providing recommendations using graph machine learning-based regression that is specifically designed to improve recommendations ... not abstract concepts but tangible/ practical improvements in the field of logistics and transportation management. ... ¶ The claimed invention, therefore, enhances existing recommendations by introducing a novel graph machine learning-based regression mechanism. ... This constitutes a clear improvement over conventional recommendations. This element explicitly integrates a human-centric consideration into what is fundamentally a technical and logistical optimization problem. ¶ ... [T]he inventive combination and explicit purpose of these functions within the claimed invention far exceed mere generic use. The claimed invention ... configures a computer to execute a highly specialized process that fundamentally transforms how delivery plans are generated and executed, leading to tangible benefits for both human operators and economic outcomes."
EXAMINER'S RESPONSE: Examiner respectfully disagrees that amended Claim 1 recites an improvement in technology. Applicant's purported "improvement over conventional recommendations," as it pertains to amended Claim 1, appears instead to recite an improvement in an abstract idea. Applicant's argued improved graph learning mechanism recites additional elements that appear merely to reflect use of computer machinery to perform the abstract ideas, mere data gathering and outputting, or generally linking the abstract idea to a field of use.
Examiner further notes that amended Claim 1 does not appear to recite or reflect "improvements in the field of logistics and transportation management," "existing recommendations," "a technical and logistical optimization problem," "delivery plans are generated and executed," or "economic outcomes." Although the claims are interpreted in light of the specification, limitations from the specification are not read into the claims.
Applicant' s arguments, see pages 14-16, filed 02 March 2026, with respect to the rejections of Claims 1-20 under 35 U.S.C. 103 have been considered but are moot because the new ground of rejection does not rely on any reference applied in the prior rejection of record for any teaching or matter specifically challenged in the argument.
APPLICANT'S ARGUMENT: Applicant argues (page 15, paragraph 3) that "Feinberg does not appear to disclose '... a total number of the set of key-phrase nodes for the researcher' (underlined emphasis added). Thus, at least the noted feature of claims 1, 16, and 20 distinguish over Feinberg."
EXAMINER'S RESPONSE: Examiner notes that Applicant's argument is moot. Amended Claim 1 is now rejected in view of Feinberg in view of Martinez Canedo in view of Ferguson.
Claim Rejections - 35 USC § 112
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claims 1, 16, 19, and 20, and dependent Claims 1-8 and 11-15, 17, and 18, are rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
Claim 1 recites the limitation "the set of key-phrase nodes for the researcher." There is insufficient antecedent basis for this limitation in the claim. Claim 1 previously recites a researcher ("an unseen subgraph associated with a first entity ... the unseen graph being ... associated with a researcher") and previously recites key-phrase nodes ("a set of key-phrase nodes of one or more first subgraphs ... associated with the third entity"), but does not appear to have previously recited a set of key-phrase nodes for the researcher, nor does it appear to have made a set of key-phrase nodes a requirement for all subgraphs of the graph database. For the purpose of examination, the Claim 1 limitation "the set of key-phrase nodes for the researcher" has been interpreted to read "all key-phrase nodes of the subgraphs of the database that are associated with the researcher." Claims 16 and 20 are rejected under the same rationale as Claim 1. Appropriate correction or clarification is required.
Claim 19 is rejected under 35 U.S.C. 112(d) or pre-AIA 35 U.S.C. 112, 4th paragraph, as being of improper dependent form for failing to further limit the subject matter of the claim upon which it depends, or for failing to include all the limitations of the claim upon which it depends. The limitations of unamended Claim 19 recite the limitations of amended Claim 16 without further limitation. Applicant may cancel the claim, amend the claim to place the claim in proper dependent form, rewrite the claim in independent form, or present a sufficient showing that the dependent claim complies with the statutory requirements.
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 1-8 and 11-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Regarding Claim 1
Step 1
Claim 1 recites a method and thus the claimed process falls within a statutory category of invention.
Step 2A Prong 1
The claim recites extract a set of subgraphs, each of which includes information associating a respective entity of an entity group with an item, which is a mental process. The claim recites computing a label score for each subgraph of the set of subgraphs, the label score being indicative of an importance of the item to the respective entity ... the label score associated with a first subgraph, associated with a third entity different from the first entity and the second entity, from the set of subgraphs is computed based on at least one of: a first score that includes a sum of nodes and edges of the first subgraph; a second score that includes an edge-to-node ratio for the first subgraph; a third score that includes a relevancy-based key-phrase score for an item associated with the first subgraph; or a fourth score that includes a relative key-phrase score for the item associated with the first subgraph, which is a mental process and/or a mathematical calculation. The claim recites generating a training dataset to include the set of subgraphs and the label score for each subgraph of the set of subgraphs, which is a mental process. The claim recites generating a prediction score as an output of the trained ML regression model for the input, which is a mental process. The claim recites determining, from the set of subgraphs, one or more subgraphs associated with the second entity, based on the prediction score, which is a mental process. The claim recites determining a ... recommendation for one or more items, based on the one or more subgraphs, which is a mental process. The claim recites the trained ML regression model being associated with a second entity different from the first entity, which is a mental process. The claim recites determining the third score by: determining a number of occurrences of a first key-phrase node in a set of key-phrase nodes of one or more first subgraphs of the set of subgraphs, the one or more first subgraphs is associated with the third entity, which is a mental process and/or a mathematical calculation. The claim recites determining a first k-score for the first key-phrase node based on a first ratio of the number of occurrences of the first key-phrase node to a total number of the set of key-phrase nodes, which is a mental process and/or a mathematical calculation. The claim recites assigning a first weight to the first k-score to determine a first weighted k score for the first key-phrase node, based on a relevancy associated with the first key-phrase node, which is a mental process and/or a mathematical calculation. The claim recites normalizing a set of weighted k-scores, including the first weighted k score for the set of key-phrase nodes, which is a mental process and/or a mathematical calculation. The claim recites summing the normalized set of weighted k-scores to determine the third score for the item associated with the first subgraph, which is a mental process and/or a mathematical calculation.
Thus, the claim recites an abstract idea.
Step 2A Prong 2
The additional element executed by a processor invokes a computer or other machinery merely as a tool to perform an existing process (see MPEP 2106.05(f), "apply it"). The additional element querying a graph database invokes a computer or other machinery merely as a tool to perform an existing process (see MPEP 2106.05(f), "apply it"). The additional element training a set of machine learning (ML) regression models on respective entity-specific subsets of the training dataset invokes a computer or other machinery merely as a tool to perform an existing process (see MPEP 2106.05(f), "apply it"). The additional element providing an unseen subgraph associated with a first entity as an input to a trained ML regression model of the trained set of ML regression models invokes a computer or other machinery merely as a tool to perform an existing process (see MPEP 2106.05(f), "apply it"). The additional element the unseen graph being a contribution graph associated with a researcher and a research journal does not amount to more than generally linking the use of a judicial exception to a particular field of use (see MPEP 2106.05(h)). The additional element a journal recommendation does not amount to more than generally linking the use of a judicial exception to a particular field of use (see MPEP 2106.05(h)). The additional element controlling a user device associated with the first entity to display the recommendation amounts to insignificant extra-solution activity (see MPEP 2106.05(g), "mere data gathering"). The additional element the set of key-phrase nodes for the researcher does not amount to more than generally linking the use of a judicial exception to a particular field of use (see MPEP 2106.05(h)).
Step 2B
The additional element executed by a processor invokes a computer or other machinery merely as a tool to perform an existing process (see MPEP 2106.05(f), "apply it"). The additional element querying a graph database invokes a computer or other machinery merely as a tool to perform an existing process (see MPEP 2106.05(f), "apply it"). The additional element training a set of machine learning (ML) regression models on respective entity-specific subsets of the training dataset invokes a computer or other machinery merely as a tool to perform an existing process (see MPEP 2106.05(f), "apply it"). The additional element providing an unseen subgraph associated with a first entity as an input to a trained ML regression model of the trained set of ML regression models invokes a computer or other machinery merely as a tool to perform an existing process (see MPEP 2106.05(f), "apply it"). The additional element the unseen graph being a contribution graph associated with a researcher and a research journal does not amount to more than generally linking the use of a judicial exception to a particular field of use (see MPEP 2106.05(h)). The additional element a journal recommendation does not amount to more than generally linking the use of a judicial exception to a particular field of use (see MPEP 2106.05(h)). The additional element controlling a user device associated with the first entity to display the recommendation is well-understood, routine, conventional activity (MPEP 2106.05(d), as supported by Freiberger, col. 16, lines 6-16: "The acquisition instructions 331 and content data update instructions 332 are generic sets of instructions that can be tailored by a content provider. The content data acquisition instructions 330 can also include user interface installation instructions 333 that enable content providers to install a user interface in the information environment (e.g., Web page) of the content provider so that users can request sets of content data from the content provider. Such user interface installation instructions are conventional and readily available for use with the attention manager of the invention"). The additional element the set of key-phrase nodes for the researcher does not amount to more than generally linking the use of a judicial exception to a particular field of use (see MPEP 2106.05(h)).
The claim lacks additional elements that integrate it into a practical application or provide significantly more, so it is directed to an abstract idea and is ineligible.
Regarding Claim 2
Step 1
Regarding Claim 2, the rejection of Claim 1 is incorporated.
Step 2A Prong 1
The claim recites querying a graph database to extract a set of subgraphs, each of which includes information associating a respective entity of an entity group with an item (as recited by Claim 1), wherein the entity corresponds to one of a user, a group of users who works together to achieve a common objective, or an organization, which is a mental process. Thus, the claim recites an abstract idea.
The claim lacks additional elements that integrate it into a practical application or provide significantly more, so it is directed to an abstract idea and is ineligible.
Regarding Claim 3
Step 1
Regarding Claim 3, the rejection of Claim 1 is incorporated.
Step 2A Prong 1
The claim recites querying a graph database to extract a set of subgraphs, each of which includes information associating a respective entity of an entity group with an item (as recited by Claim 1), wherein the graph database includes a knowledge graph which comprises of: a set of distinct entity nodes associated with the entity group, a set of distinct item nodes corresponding to a set of items, and the set of subgraphs, each of which lies between at least one entity node of the set of distinct entity nodes and at least one item node of the set of distinct item nodes, which is a mental process. Thus, the claim recites an abstract idea.
Step 2A Prong 2, Step 2B
The claim lacks additional elements that integrate it into a practical application or provide significantly more, so it is directed to an abstract idea and is ineligible.
Regarding Claim 4
Step 1
Regarding Claim 4, the rejection of Claim 1 is incorporated.
Step 2A Prong 1
Claim 4 recites the abstract ideas recited by parent Claim 1.
Step 2A Prong 2, Step 2B
The additional element training a set of machine learning (ML) regression models on respective entity-specific subsets of the training dataset (as recited by Claim 1), wherein each ML regression model of the set of ML regression models is a graph ML model does not amount to more than generally linking the use of a judicial exception to a particular field of use (see MPEP 2106.05(h)).
The claim lacks additional elements that integrate it into a practical application or provide significantly more, so it is directed to an abstract idea and is ineligible.
Regarding Claim 5
Step 1
Regarding Claim 5, the rejection of Claim 1 is incorporated.
Step 2A Prong 1
The claim recites querying a graph database to extract a set of subgraphs, each of which includes information associating a respective entity of an entity group with an item (as recited by Claim 1), wherein the information in each subgraph of the set of subgraphs comprises of a set of intermediate nodes connected to an entity node and an item node through a set of edges, and each intermediate node of the set of intermediate nodes represents an attribute that is linked to the item, which is a mental process. Thus, the claim recites an abstract idea.
Step 2A Prong 2, Step 2B
The claim lacks additional elements that integrate it into a practical application or provide significantly more, so it is directed to an abstract idea and is ineligible.
Regarding Claim 6
Step 1
Regarding Claim 6, the rejection of Claim 5 is incorporated.
Step 2A Prong 1
The claim recites querying a graph database to extract a set of subgraphs, each of which includes information associating a respective entity of an entity group with an item (as recited by Claim 1), wherein the information in each subgraph of the set of subgraphs comprises of a set of intermediate nodes connected to an entity node and an item node through a set of edges, and wherein each intermediate node of the set of intermediate nodes represents an attribute that is linked to the item (as recited by Claim 5), wherein the respective entity of the entity group corresponds to the researcher and the item corresponds to the research journal with which the researcher is associated, wherein the set of intermediate nodes include a set of key-phrase nodes and a set of topic nodes, which is a mental process. Thus, the claim recites an abstract idea.
Step 2A Prong 2, Step 2B
The claim lacks additional elements that integrate it into a practical application or provide significantly more, so it is directed to an abstract idea and is ineligible.
Regarding Claim 7
Step 1
Regarding Claim 7, the rejection of Claim 6 is incorporated.
Step 2A Prong 1
The claim recites wherein a relevancy weight is assigned to an edge between each key-phrase node of the set of key-phrase nodes and a corresponding topic node of the set of topic nodes, which is a mental process. Thus, the claim recites an abstract idea.
Step 2A Prong 2, Step 2B
The claim lacks additional elements that integrate it into a practical application or provide significantly more, so it is directed to an abstract idea and is ineligible.
Regarding Claim 8
Step 1
Regarding Claim 8, the rejection of Claim 1 is incorporated.
Step 2A Prong 1
The claim recites determining a set of distinct entity nodes in the graph database, which is a mental process. The claim recites determining a set of distinct item nodes in the graph database, which is a mental process. The claim recites performing, for each pair of an entity node of the set of distinct entity nodes and an item node of the set of distinct item nodes, a set of operations, which is a mental process. The claim recites a first operation to extract a first subgraph of a plurality of subgraphs, wherein the information in the extracted first subgraph includes: a set of intermediate nodes, and a set of edges, each connecting a pair of intermediate nodes of the set of intermediate nodes, which is a mental process. The claim recites a second operation to trim one or more intermediate nodes from the set of intermediate nodes, based on a relevancy attribute assigned to each edge of the set of edges, which is a mental process. The claim recites extract the set of subgraphs from the plurality of subgraphs by removing one or more duplicate subgraphs from the plurality of subgraphs, which is a mental process. Thus, the claim recites an abstract idea.
Step 2A Prong 2, Step 2B
The claim lacks additional elements that integrate it into a practical application or provide significantly more, so it is directed to an abstract idea and is ineligible.
Regarding Claim 11
Step 1
Regarding Claim 11, the rejection of Claim 1 is incorporated.
Step 2A Prong 1
The claim recites determining the fourth score by: sorting a set of k-scores for a set of key-phrase nodes of one or more first subgraphs of the set of subgraphs to select a set of top k-scores, the one or more first subgraphs being associated with the third entity, which is a mental process and/or a mathematical calculation. The claim recites determining a sum of the set of top k-scores and a sum of the set of k- scores, which is a mental process and/or a mathematical calculation. The claim recites determining a second ratio of the sum of the set of top k-scores to the sum of the set of k-scores to determine the fourth score, which is a mental process and/or a mathematical calculation. Thus, the claim recites an abstract idea.
Step 2A Prong 2, Step 2B
The claim lacks additional elements that integrate it into a practical application or provide significantly more, so it is directed to an abstract idea and is ineligible.
Regarding Claim 12
Step 1
Regarding Claim 12, the rejection of Claim 1 is incorporated.
Step 2A Prong 1
The claim recites the computation of the label score comprises of: normalizing each of the first score, the second score, the third score, and the fourth score, which is a mental process and/or a mathematical calculation. The claim recites summing each of the normalized first score, the normalized second score, the normalized third score, and the normalized fourth score, which is a mental process and/or a mathematical calculation. Thus, the claim recites an abstract idea.
Step 2A Prong 2, Step 2B
The claim lacks additional elements that integrate it into a practical application or provide significantly more, so it is directed to an abstract idea and is ineligible.
Regarding Claim 13
Step 1
Regarding Claim 13, the rejection of Claim 1 is incorporated.
Step 2A Prong 1
The claim recites partitioning the training dataset into the respective entity-specific subsets of the training dataset, which is a mental process. The claim recites cross-validating each trained ML regression model of the set of ML regression models on at least one subset of the entity-specific subsets that is not used in the training of the respective ML regression model, which is a mental process. The claim recites determining a set of adjusted scores associated with the set of subgraphs, based on the cross-validation, which is a mental process. Thus, the claim recites an abstract idea.
Step 2A Prong 2, Step 2B
The claim lacks additional elements that integrate it into a practical application or provide significantly more, so it is directed to an abstract idea and is ineligible.
Regarding Claim 14
Step 1
Regarding Claim 14, the rejection of Claim 13 is incorporated.
Step 2A Prong 1
The claim recites the one or more subgraphs associated with the second entity are determined by: sorting a first subset of adjusted scores of the set of adjusted scores associated with the second entity into a list, which is a mental process and/or a mathematical calculation. The claim recites updating the list of the first subset of adjusted scores by indexing the prediction score for the unseen subgraph in the list, which is a mental process. The claim recites comparing the prediction score in the updated list with each adjusted score of the first subset of adjusted scores, wherein the one or more subgraphs are determined based on the comparison, which is a mental process. Thus, the claim recites an abstract idea.
Step 2A Prong 2, Step 2B
The claim lacks additional elements that integrate it into a practical application or provide significantly more, so it is directed to an abstract idea and is ineligible.
Regarding Claim 15
Step 1
Regarding Claim 15, the rejection of Claim 1 is incorporated.
Step 2A Prong 1
The claim recites wherein the one or more items in the recommendation are determined based on an association of the one or more items with the one or more subgraphs, which is a mental process. Thus, the claim recites an abstract idea.
Step 2A Prong 2, Step 2B
The claim lacks additional elements that integrate it into a practical application or provide significantly more, so it is directed to an abstract idea and is ineligible.
Regarding Claim 16
Step 1
Claim 16 recites one or more non-transitory computer-readable storage media configured to store instructions that, in response to being executed, cause a system to perform operations, and thus the claimed manufacture falls within a statutory category of invention.
Step 2A Prong 1
The claim recites extract a set of subgraphs, each of which includes information associating a respective entity of an entity group with an item, which is a mental process. The claim recites computing a label score for each subgraph of the set of subgraphs, the label score being indicative of an importance of the item to the respective entity ... the label score associated with a first subgraph, associated with a third entity different from the first entity and the second entity, from the set of subgraphs is computed based on at least one of: a first score that includes a sum of nodes and edges of the first subgraph; a second score that includes an edge-to-node ratio for the first subgraph; a third score that includes a relevancy-based key-phrase score for an item associated with the first subgraph; or a fourth score that includes a relative key-phrase score for the item associated with the first subgraph, which is a mental process and/or a mathematical calculation. The claim recites generating a training dataset to include the set of subgraphs and the label score for each subgraph of the set of subgraphs, which is a mental process. The claim recites generating a prediction score as an output of the trained ML regression model for the input, which is a mental process. The claim recites determining, from the set of subgraphs, one or more subgraphs associated with the second entity, based on the prediction score, which is a mental process. The claim recites determining a ... recommendation for one or more items, based on the one or more subgraphs, which is a mental process. The claim recites the trained ML regression model being associated with a second entity different from the first entity, which is a mental process. The claim recites determining the third score by: determining a number of occurrences of a first key-phrase node in a set of key-phrase nodes of one or more first subgraphs of the set of subgraphs, the one or more first subgraphs is associated with the third entity, which is a mental process and/or a mathematical calculation. The claim recites determining a first k-score for the first key-phrase node based on a first ratio of the number of occurrences of the first key-phrase node to a total number of the set of key-phrase nodes, which is a mental process and/or a mathematical calculation. The claim recites assigning a first weight to the first k-score to determine a first weighted k score for the first key-phrase node, based on a relevancy associated with the first key-phrase node, which is a mental process and/or a mathematical calculation. The claim recites normalizing a set of weighted k-scores, including the first weighted k score for the set of key-phrase nodes, which is a mental process and/or a mathematical calculation. The claim recites summing the normalized set of weighted k-scores to determine the third score for the item associated with the first subgraph, which is a mental process and/or a mathematical calculation.
Thus, the claim recites an abstract idea.
Step 2A Prong 2
The additional element querying a graph database invokes a computer or other machinery merely as a tool to perform an existing process (see MPEP 2106.05(f), "apply it"). The additional element training a set of machine learning (ML) regression models on respective entity-specific subsets of the training dataset invokes a computer or other machinery merely as a tool to perform an existing process (see MPEP 2106.05(f), "apply it"). The additional element providing an unseen subgraph associated with a first entity as an input to a trained ML regression model of the trained set of ML regression models invokes a computer or other machinery merely as a tool to perform an existing process (see MPEP 2106.05(f), "apply it"). The additional element the unseen graph being a contribution graph associated with a researcher and a research journal does not amount to more than generally linking the use of a judicial exception to a particular field of use (see MPEP 2106.05(h)). The additional element a journal recommendation does not amount to more than generally linking the use of a judicial exception to a particular field of use (see MPEP 2106.05(h)). The additional element controlling a user device associated with the first entity to display the recommendation amounts to insignificant extra-solution activity (see MPEP 2106.05(g), "mere data gathering"). The additional element the set of key-phrase nodes for the researcher does not amount to more than generally linking the use of a judicial exception to a particular field of use (see MPEP 2106.05(h)).
Step 2B
The additional element querying a graph database invokes a computer or other machinery merely as a tool to perform an existing process (see MPEP 2106.05(f), "apply it"). The additional element training a set of machine learning (ML) regression models on respective entity-specific subsets of the training dataset invokes a computer or other machinery merely as a tool to perform an existing process (see MPEP 2106.05(f), "apply it"). The additional element providing an unseen subgraph associated with a first entity as an input to a trained ML regression model of the trained set of ML regression models invokes a computer or other machinery merely as a tool to perform an existing process (see MPEP 2106.05(f), "apply it"). The additional element the unseen graph being a contribution graph associated with a researcher and a research journal does not amount to more than generally linking the use of a judicial exception to a particular field of use (see MPEP 2106.05(h)). The additional element a journal recommendation does not amount to more than generally linking the use of a judicial exception to a particular field of use (see MPEP 2106.05(h)). The additional element controlling a user device associated with the first entity to display the recommendation is well-understood, routine, conventional activity (MPEP 2106.05(d), as supported by Freiberger, col. 16, lines 6-16: "The acquisition instructions 331 and content data update instructions 332 are generic sets of instructions that can be tailored by a content provider. The content data acquisition instructions 330 can also include user interface installation instructions 333 that enable content providers to install a user interface in the information environment (e.g., Web page) of the content provider so that users can request sets of content data from the content provider. Such user interface installation instructions are conventional and readily available for use with the attention manager of the invention"). The additional element the set of key-phrase nodes for the researcher does not amount to more than generally linking the use of a judicial exception to a particular field of use (see MPEP 2106.05(h)).
The claim lacks additional elements that integrate it into a practical application or provide significantly more, so it is directed to an abstract idea and is ineligible.
Claims 17-19, dependent on Claim 16, inherit the rejection of Claim 16. Claims 17-19 incorporate substantively all limitations of Claims 3, 8, and 9, respectively, in non-transitory computer-readable storage media form, and are rejected under the same rationales.
Regarding Claim 20
Step 1
Claim 20 recites a system, comprising: a memory configured to store instructions; and a processor, coupled to the memory, executes the instructions to perform a process, and thus the claimed machine falls within a statutory category of invention.
Step 2A Prong 1
The claim recites extract a set of subgraphs, each of which includes information associating a respective entity of an entity group with an item, which is a mental process. The claim recites computing a label score for each subgraph of the set of subgraphs, the label score being indicative of an importance of the item to the respective entity ... the label score associated with a first subgraph, associated with a third entity different from the first entity and the second entity, from the set of subgraphs is computed based on at least one of: a first score that includes a sum of nodes and edges of the first subgraph; a second score that includes an edge-to-node ratio for the first subgraph; a third score that includes a relevancy-based key-phrase score for an item associated with the first subgraph; or a fourth score that includes a relative key-phrase score for the item associated with the first subgraph, which is a mental process and/or a mathematical calculation. The claim recites generating a training dataset to include the set of subgraphs and the label score for each subgraph of the set of subgraphs, which is a mental process. The claim recites generating a prediction score as an output of the trained ML regression model for the input, which is a mental process. The claim recites determining, from the set of subgraphs, one or more subgraphs associated with the second entity, based on the prediction score, which is a mental process. The claim recites determining a ... recommendation for one or more items, based on the one or more subgraphs, which is a mental process. The claim recites the trained ML regression model being associated with a second entity different from the first entity, which is a mental process. The claim recites determining the third score by: determining a number of occurrences of a first key-phrase node in a set of key-phrase nodes of one or more first subgraphs of the set of subgraphs, the one or more first subgraphs is associated with the third entity, which is a mental process and/or a mathematical calculation. The claim recites determining a first k-score for the first key-phrase node based on a first ratio of the number of occurrences of the first key-phrase node to a total number of the set of key-phrase nodes, which is a mental process and/or a mathematical calculation. The claim recites assigning a first weight to the first k-score to determine a first weighted k score for the first key-phrase node, based on a relevancy associated with the first key-phrase node, which is a mental process and/or a mathematical calculation. The claim recites normalizing a set of weighted k-scores, including the first weighted k score for the set of key-phrase nodes, which is a mental process and/or a mathematical calculation. The claim recites summing the normalized set of weighted k-scores to determine the third score for the item associated with the first subgraph, which is a mental process and/or a mathematical calculation.
Thus, the claim recites an abstract idea.
Step 2A Prong 2
The additional element querying a graph database invokes a computer or other machinery merely as a tool to perform an existing process (see MPEP 2106.05(f), "apply it"). The additional element training a set of machine learning (ML) regression models on respective entity-specific subsets of the training dataset invokes a computer or other machinery merely as a tool to perform an existing process (see MPEP 2106.05(f), "apply it"). The additional element providing an unseen subgraph associated with a first entity as an input to a trained ML regression model of the trained set of ML regression models invokes a computer or other machinery merely as a tool to perform an existing process (see MPEP 2106.05(f), "apply it"). The additional element the unseen graph being a contribution graph associated with a researcher and a research journal does not amount to more than generally linking the use of a judicial exception to a particular field of use (see MPEP 2106.05(h)). The additional element a journal recommendation does not amount to more than generally linking the use of a judicial exception to a particular field of use (see MPEP 2106.05(h)). The additional element controlling a user device associated with the first entity to display the recommendation amounts to insignificant extra-solution activity (see MPEP 2106.05(g), "mere data gathering"). The additional element the set of key-phrase nodes for the researcher does not amount to more than generally linking the use of a judicial exception to a particular field of use (see MPEP 2106.05(h)).
Step 2B
The additional element querying a graph database invokes a computer or other machinery merely as a tool to perform an existing process (see MPEP 2106.05(f), "apply it"). The additional element training a set of machine learning (ML) regression models on respective entity-specific subsets of the training dataset invokes a computer or other machinery merely as a tool to perform an existing process (see MPEP 2106.05(f), "apply it"). The additional element providing an unseen subgraph associated with a first entity as an input to a trained ML regression model of the trained set of ML regression models invokes a computer or other machinery merely as a tool to perform an existing process (see MPEP 2106.05(f), "apply it"). The additional element the unseen graph being a contribution graph associated with a researcher and a research journal does not amount to more than generally linking the use of a judicial exception to a particular field of use (see MPEP 2106.05(h)). The additional element a journal recommendation does not amount to more than generally linking the use of a judicial exception to a particular field of use (see MPEP 2106.05(h)). The additional element controlling a user device associated with the first entity to display the recommendation is well-understood, routine, conventional activity (MPEP 2106.05(d), as supported by Freiberger, col. 16, lines 6-16: "The acquisition instructions 331 and content data update instructions 332 are generic sets of instructions that can be tailored by a content provider. The content data acquisition instructions 330 can also include user interface installation instructions 333 that enable content providers to install a user interface in the information environment (e.g., Web page) of the content provider so that users can request sets of content data from the content provider. Such user interface installation instructions are conventional and readily available for use with the attention manager of the invention"). The additional element the set of key-phrase nodes for the researcher does not amount to more than generally linking the use of a judicial exception to a particular field of use (see MPEP 2106.05(h)).
The claim lacks additional elements that integrate it into a practical application or provide significantly more, so it is directed to an abstract idea and is ineligible.
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.
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.
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.
Claims 1-6, 8, 12, and 15-20 are rejected under 35 U.S.C. 103 as being unpatentable over Feinberg, et al. (US 2014/0280224 A1, hereinafter "Feinberg") in view of Martinez Canedo, et al. (US 2019/0095806 A1, hereinafter "Canedo") in view of Ferguson, et al., "An Investigation of Term Weighting Approaches for Microblog Retrieval" (hereinafter "Ferguson").
Regarding Claim 1, Feinberg recites:
A method, executed by a processor (Feinberg, [0002]: "The present invention relates to systems and methods for augmenting graph databases and more specifically to the identification of relationships that can be used to augment a graph database"), comprising:
querying a graph database (Feinberg, [0045]: "the client devices are configured to communicate with the relationship recommendation server system 110 to request and obtain recommended relationships ... for a particular graph database representing a collaborative environment") to extract a set of subgraphs (Feinberg, [0062]: "The search node is utilized to identify ... recommended relationships. ... [T]hese recommended relationships result in the inclusion of one or more subgraphs"), each of which includes information associating a respective entity of an entity group with an item (Feinberg, [0036]: "The academic publications and products created by these teams typically have multiple authors and other collaborators. ... [F]or a graph database representing a research environment, the nodes can represent individual researchers while the edges represent collaboration (e.g. co-authorship) between researchers," where Feinberg's collaborators, individual researcher, and academic publication correspond to the instant entity group, entity, and item);
computing a label score for each subgraph of the set of subgraphs, the label score being indicative of an importance of the item to the respective entity (Feinberg, [0006]: "the graph database includes a set of nodes including node attribute data and a set of edges including edge attribute data and edges in the set of edges describe relationships between nodes in the set of nodes ... obtain a search node, generate additional edges between the search node and a subset of the nodes within the graph database" and [0016]: "the identification of a subgraph included in the set of subgraphs is based on the similarity score and a statistical significance of the subgraph within the graph database," where Feinberg's similarity score corresponds to the instant label score); ... ;
providing an unseen subgraph associated with a first entity (Feinberg, [0062]: "From the perspective of the search node, the graph database is augmented with edges and the resulting subgraphs created including the augmented edge," where Feinberg's created subgraphs corresponds to the instant unseen subgraph and Feinberg's augmented with edges corresponds to the instant associated) as an input to a ... regression model (Feinberg, [0038]: "the keywords and concepts are mapped to the attributes describing the nodes and edges within the graph database in order to identify the relevant subgraphs. For example ... according to a similarity scoring function," where Feinberg's scoring function corresponds to the instant regression model, as in the instant specification at [00039]: "a set of regression models, each of which may be trained to identify a relationship between inputs, such as features in a training dataset and to predict scores that indicate importance of items"), ... the unseen graph being a contribution graph associated with a researcher and a research journal (Feinberg, [0039]: "a similarity score is utilized to identify subgraphs and motifs within the graph database that are isomorphic to each other.... The relationship recommendation system can then ... determine a set of edges between researchers .... Success metrics include ... successful academic publications, citations accrued");
generating a prediction (Feinberg, [0038]: "the relationship recommendation system can ... determine a 'team' of nodes ... that will be most likely to form a productive solution") score as an output of the... regression model for the input (Feinberg, [0062]: "From the perspective of the search node, the graph database is augmented with edges and the resulting subgraphs created including the augmented edge are ranked according to the identified motifs. ... Once the subgraphs have been generated by adding edges ... a ranking function ... can be produced ... to provide recommended relationships for a particular concept description," where Feinberg's ranking corresponds to the instant prediction score);
determining, from the set of subgraphs, one or more subgraphs associated with the second entity, based on the prediction score (Feinberg, [0052]: "Based on the attributes of the nodes 410 and edges 420 within a particular motif, a similarity score can be determined. In many embodiments, the similarity score is also based on a particular concept description that is to be solved by suggesting nodes and/or edges based on the identified motifs. In this way, the motifs can be tailored to identify beneficial relationships for a particular problem. The similarity score can be utilized to compare and rank subgraphs occurring within the graph database");
determining a journal recommendation for one or more items, based on the one or more subgraphs (Feinberg, [0039]: "Success metrics include ... successful academic publications .... The relationship recommendation system can ... suggest the nodes or edges (e.g. the ... collaborations) that would result in the desired combination ... being added to the graph database"); and
controlling a user device associated with the first entity to display the recommendation (Feinberg, [0042]: "a scientist could write an abstract describing a proposed drug delivery method, and the relationship recommendation system would output a list of scientists, physicians, and regulatory experts" and [0045]: "Client devices include, but are not limited to, mobile devices 116, personal computers 118, and interactive platforms"),
the label score associated with a first subgraph, associated with a third entity different from the first entity and the second entity, from the set of subgraphs (Feinberg, [0052]: "Based on the attributes of the nodes 410 and edges 420 within a particular motif, a similarity score can be determined. ... The similarity score can be utilized to compare and rank subgraphs occurring within the graph database. In this way, motifs identified as having positive connotations can be identified within those subgraphs, including those subgraphs that are isomorphic to a reference motif") is computed based on at least one of:
a first score that includes a sum of nodes and edges of the first subgraph (Feinberg, [0054]: "alternative functions can be utilized to compute similarity scores for any two sets of nodes from structurally isomorphic subgraphs .... [T]he similarity score can be framed as a discrete optimization program by calculating a similarity score pairwise by node with
n
i
and
n
j
corresponding to nodes in the two subsets N 1 and N 2 for the subgraphs chosen without replacement: [equation] ... In this example, the function
d
a
is minimized over all possible choices of
g
that maps a node in
N
1
to a node in
N
2
," where Feinberg's depicted
d
a
calculates sums over the nodes and edges of the subgraphs, reasonably suggesting a similarity to a first subgraph);
a second score that includes an edge-to-node ratio for the first subgraph (Feinberg, [0070]: "it is common in academia for particular researchers to be extremely well connected with other researchers .... By utilizing representative nodes corresponding to these well-connected nodes (e.g. popular researchers) ... additional recommended relationships can be determined based the search node and those nodes connected to the representative node," where Feinberg's similarity based on well-connected nodes reasonably suggests a score based on the high edge-to-node ratio of the well-connected node);
a third score that includes a relevancy-based key-phrase score for an item associated with the first subgraph (Feinberg, [0070]: "one attribute could be a word frequency count among academic publications published by a research group .... This augmented recommended relationship can be directly between the search node and the additional nodes"); or
a fourth score that includes a relative key-phrase score for the item associated with the first subgraph (Feinberg, [0038]: "the keywords and concepts are mapped to the attributes describing the nodes and edges within the graph database in order to identify the relevant subgraphs. For example, the relationship recommendation system can be employed ... to determine a 'team' of nodes ... that will be most likely to form a productive solution to the concept description," where Feinberg's most relevant subgraphs identified using keywords corresponds to the instant includes relative key-phrase score);
determining the third score by:
determining a number of occurrences of a first key-phrase node in a set of key-phrase nodes of one or more first subgraphs of the set of subgraphs, the one or more first subgraphs is associated with the third entity (Feinberg, [0070]: "one attribute could be a word frequency count among academic publications published by a research group ... By utilizing representative nodes corresponding to these well-connected nodes (e.g. popular researchers), when a recommended relationship between the representative node and the search node is identified, additional recommended relationships can be determined based the search node and those nodes connected to the representative node," where Feinberg's search node corresponds to the instant first subgraph and Feinberg's research group corresponds to the instant third entity);
determining a first k-score for the first key-phrase node ... for the researcher (Feinberg, [0070]: "one attribute could be a word frequency count among academic publications published by a research group ... By utilizing representative nodes corresponding to these well-connected nodes (e.g. popular researchers), when a recommended relationship between the representative node and the search node is identified, additional recommended relationships can be determined based the search node and those nodes connected to the representative node," where Feinberg's search node corresponds to the instant researcher, given that Feinberg has previously indicated similarity being useful to identify research collaborators for a researcher, as in [0039]: "The relationship recommendation system can then identify which patterns of researchers tend to collaborate, identify those collaborations that are productive, and determine a set of edges between researchers based on those motifs within the graph database indicative of a likelihood of a fruitful collaboration, measure by a success metric");
assigning a first weight to the first k-score to determine a first weighted k score for the first key-phrase node, based on a relevancy associated with the first key-phrase node (Feinberg, [0053]: "one can define the similarity score
d
for attribute a as: ...
d
M
1
,
M
2
=
∑
a
c
a
d
a
M
1
,
M
2
2
1
2
where
c
a
is a constant for attribute
a
. In a number of embodiments,
c
a
is a constant utilized to weight attributes related to the nodes and edges within the motifs," where Feinberg's
c
a
and
d
a
correspond to the instant weight and k-score, respectively, given that Feinberg indicates that keywords may be node and edge attributes, as in [0038]: "Concept description data can be provided that includes keywords and concepts ... mapped to the attributes describing the nodes and edges within the graph database in order to identify the relevant subgraphs");
normalizing a set of weighted k-scores, including the first weighted k score for the set of key-phrase nodes; and summing the normalized set of weighted k-scores to determine the third score for the item associated with the first subgraph (Feinberg, [0053]: "a similarity score
d
for attribute
a
for the motifs can be defined by various metrics and norms. As one example, one can define the similarity score
d
for attribute a as: ...
d
M
1
,
M
2
=
∑
a
c
a
d
a
M
1
,
M
2
2
1
2
," where Feinberg's
d
M
1
,
M
2
is calculated from a norm of the weighted
d
a
scores ).
Feinberg teaches using a graph database for recommending novel relationships between research entities based on similarity scores calculated using a regression model defined over the nodes and edges of subgraphs of the graph.
Feinberg does not explicitly teach generating a training dataset to include the set of subgraphs and the label score for each subgraph of the set of subgraphs; training a set of machine learning (ML) regression models on respective entity-specific subsets of the training dataset; and providing an unseen subgraph associated with a first entity as an input to a trained ML regression model of the trained set of ML regression models, the trained ML regression model being associated with a second entity different from the first entity.
However, Canedo teaches:
generating a training dataset to include the set of subgraphs and the label score for each subgraph of the set of subgraphs (Canedo, [0033]: "During the training process of the SGCNN, the labels of the functional graph examples are either annotated by domain experts, or are automatically extracted from the engineering software. ... Functional graphs are different; this means that the subgraph representing system A is different than the subgraph representing system B. Therefore, the SGCNN architecture trains different GCNNs based on the functional graph properties");
training a set of machine learning (ML) regression models on respective entity-specific subsets of the training dataset (Canedo, [0034]: "the complete graph (or part of it) is provided as an input for each hidden layer of the neural networks to guide the training process and help the model to converge faster. The context is provided by the location of these functional graphs in the complete graph. After training, the generated predictors ingest a subgraph and its context, and classify it to the functional labels" and [0062]: "the neighbor node feature vectors may be derived for each subgraph by generating a neighbor feature matrix comprising neighbor paths in the subgraph, and then applying a 2D convolution operation with a trainable matrix and a bias variable to extract the neighbor node feature vectors from the neighbor feature matrix. The neighbor feature matrix may be generated for each subgraph by first determining all neighboring paths for the subgraph");
providing an unseen subgraph associated with a first entity as an input to a trained ML regression model of the trained set of ML regression models, the trained ML regression model being associated with a second entity different from the first entity (Canedo, [0034]: "After training, the generated predictors ingest a subgraph and its context, and classify it to the functional labels").
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the teachings of Feinberg regarding Feinberg teaches using a graph database for recommending novel relationships between research entities based on similarity scores calculated using a regression model defined over the nodes and edges of subgraphs of the graph with those of Canedo regarding generating a training dataset to include the set of subgraphs and the label score for each subgraph of the set of subgraphs; training a set of machine learning (ML) regression models on respective entity-specific subsets of the training dataset; and providing an unseen subgraph associated with a first entity as an input to a trained ML regression model of the trained set of ML regression models, the trained ML regression model being associated with a second entity different from the first entity.
The motivation to do so would be to facilitate training of neural network models that provide improved labeling performance over irregular graph structures and features (Canedo, [0029]: "One shortcoming of the initial incarnation of GCNNs is that they can only handle graphs with the same number of nodes and with similar graph structures. ... Some subset of these graph instances are labelled at the node level. One can train a GCNN which, given a new graph instance, will provide a node-level labelling based on what it has observed .... in order to better label nodes in the graph").
The Feinberg/Canedo combination teaches determining a first k-score for the first key-phrase node for the researcher.
The Feinberg/Canedo combination does not explicitly teach determining a first k-score for the first key-phrase node based on a first ratio of the number of occurrences of the first key-phrase node to a total number of the set of key-phrase nodes.
However Ferguson teaches:
determining a first k-score for the first key-phrase node based on a first ratio of the number of occurrences of the first key-phrase node to a total number of the set of key-phrase nodes (Ferguson, p. 553, 2 Ranking Microblogs: "Standard information retrieval ranking algorithms ... use inverse document frequency statistics to downweight non-discrimimnative terms that occur in many documents. The Okapi BM25 model ... computes the similarity of a document d to query q, containing the terms t, as: [Eq. (1)]," where Ferguson's Eq. 1 depicts a ratio of a number of occurrences of a term in a query (
t
f
i
) to a total number of occurrences of a term (
d
f
i
)).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the teachings of the Feinberg/Canedo combination regarding determining a first k-score for the first key-phrase node for the researcher with those of Ferguson regarding determining a first k-score for the first key-phrase node based on a first ratio of the number of occurrences of the first key-phrase node to a total number of the set of key-phrase nodes.
The motivation to do so would be to facilitate retrieval based on more discriminative terms (Ferguson, p. 553, 2 Ranking Microblogs: "Standard information retrieval ranking algorithms ... use inverse document frequency statistics to downweight non-discrimimnative terms that occur in many documents. ... The BM25 model has been extensively tested on the TREC test collections, and has consistently been shown to perform effectively. For this reason, and because it allows for the adjustment of the term frequency weighting and document length normalisation through the ... parameters").
Regarding Claim 16, Feinberg teaches:
One or more non-transitory computer-readable storage media configured to store instructions that, in response to being executed, cause a system to perform operations (Feinberg, Claim 1: "A relationship recommendation server system, comprising: a processor; and a memory connected to the processor and configured to store a relationship recommendation program"), the operations comprising: precisely those steps recited by the method of Claim 1. Claim 20 is rejected under the same rationale.
Regarding Claim 20, Feinberg teaches:
A system, comprising: a memory configured to store instructions; and a processor, coupled to the memory, executes the instructions to perform a process (Feinberg, [0006]: "a relationship recommendation server system includes a processor and a memory connected to the processor and configured to store a relationship recommendation program, wherein the relationship recommendation program configures the processor to obtain a graph database") comprising: precisely those steps recited by the method of Claim 1. Claim 20 is rejected under the same rationale.
Regarding Claim 2, the rejection of Claim 1 is incorporated. The Feinberg/Canedo/Ferguson combination teaches:
wherein the entity corresponds to one of a user, a group of users who works together to achieve a common objective, or an organization (Feinberg, [0036]: "Networks of collaborators are also becoming increasingly important in the field of biomedicine, as scientists have increasingly mapped protein and gene interaction networks, forming the interactome for various species. The academic publications and products created by these teams typically have multiple authors and other collaborators. ... [F]or a graph database representing a research environment, the nodes can represent individual researchers while the edges represent collaboration ( e.g. co-authorship) between researchers," where Feinberg's individual researcher corresponds to the instant user).
Regarding Claim 3, the rejection of Claim 1 is incorporated. The Feinberg/Canedo/Ferguson combination teaches:
wherein the graph database includes a knowledge graph (Feinberg, [0044]: "the relationship recommendation server system 110 is configured to obtain source data describing one or more collaborative environments from an information source system 112 and generate (or augment) a graph database based on the obtained source data. The source data includes a variety of data, including entity information and relationship information along with any other metadata describing the collaborative environment, that can be utilized by the relationship recommendation server system 110 to generate nodes (representing the entities) and edges (representing the relationships) for inclusion in the graph database," where Feinberg's source data for collaborative research environments corresponds to the instant knowledge graph, per the instant specification at [00002]: "One of the common types of graph is a knowledge graph that represents objects, entities, events, situations, and associated interdependencies in the form of a graph-structured data model") which comprises of:
a set of distinct entity nodes associated with the entity group (Feinberg, [0036]: "for a graph database representing a research environment, the nodes can represent individual researchers while the edges represent collaboration ( e.g. co-authorship) between researchers" and [0038]: "this team of nodes can represent a team of researchers likely to have a productive collaboration for a particular research project," where Feinberg's researchers are distinct due to membership in a controlled set of entities, as in: [0024]: "the method further includes obtaining source data using the relationship recommendation server system, where the source data describes a collaborative environment including a set of entities and a set of relationships describing the relationships between the entities in the set of entities and generating the obtained graph database based on the obtained source data using the relationship recommendation server system"),
a set of distinct item nodes corresponding to a set of items (Feinberg, [0036]: "The academic publications and products created by these teams typically have multiple authors and other collaborators. ... [F]or a graph database representing a research environment, the nodes can represent individual researchers while the edges represent collaboration ( e.g. co-authorship) between researchers" and [0037]: "research networks can spawn directed citation graphs where nodes are either authors or papers who have referenced each other and professional networks can spawn graphs of professionals who have collaborated on projects or worked in the same groups at companies," where Feinberg's academic publications corresponds to the instant set of items and where Feinberg's academic publications are also entities included in a set of entities, as in [0024]: "the source data describes a collaborative environment including a set of entities ... and generating the obtained graph database based on the obtained source data" and [0037]: "A graph database for a particular collaborative environment can represent a variety of relationships between entities. For example ... research networks can spawn directed citation graphs where nodes are either authors or papers who have referenced each other"), and
the set of subgraphs, each of which lies between at least one entity node of the set of distinct entity nodes and at least one item node of the set of distinct item nodes (Feinberg, [0036]: "for a graph database representing a research environment, the nodes can represent individual researchers while the edges represent collaboration (e.g. co-authorship) between researchers," where Fienberg's nodes representing individual researchers corresponds to the instant entity subgraph and where Fienberg's research networks comprise co-authorship edges between individual researchers and their collaborators, as in [0037]: "research networks can spawn directed citation graphs where nodes are either authors or papers").
Claim 17 incorporates substantively all limitations of Claim 3 in non-transitory computer-readable storage media form and is rejected under the same rationale as Claim 3.
Regarding Claim 4, the rejection of Claim 1 is incorporated. Canedo further teaches:
wherein each ML regression model of the set of ML regression models is a graph ML model (Canedo, [0034]: "After training, the generated predictors ingest a subgraph and its context, and classify it to the functional labels" and [0068]: "the SGCNN is used to determine functional scores for the system. Each functional score corresponds to one of a plurality functional labels describing the subsystems," where Canedo's scoring predictors are regression models, as in the instant specification at [00039]: "a set of regression models, each of which may be trained to identify a relationship between inputs, such as features in a training dataset and to predict scores that indicate importance of items").
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the prior teachings of the Feinberg/Canedo/Ferguson combination regarding a set of models of the training dataset with the further teachings of Canedo regarding each ML regression model of the set of ML regression models being a graph ML model.
The motivation to do so would be to facilitate training of neural network models that provide improved labeling performance over irregular graph structures and features (Canedo, [0029]: "One shortcoming of the initial incarnation of GCNNs is that they can only handle graphs with the same number of nodes and with similar graph structures. ... Some subset of these graph instances are labelled at the node level. One can train a GCNN which, given a new graph instance, will provide a node-level labelling based on what it has observed .... in order to better label nodes in the graph").
Regarding Claim 5, the rejection of Claim 1 is incorporated. The Feinberg/Canedo/Ferguson combination teaches:
wherein the information in each subgraph of the set of subgraphs comprises of a set of intermediate nodes connected to an entity node and an item node through a set of edges (Feinberg, [0017]: "the relationship recommendation application further configures the processor to rank the subgraphs generated based on the search node and the additional edges and recommend those relationships corresponding to the additional edges forming subgraphs have a rank exceeding a threshold value," where Feinberg's generated subgraphs corresponds to the instant intermediate nodes), and
each intermediate node of the set of intermediate nodes represents an attribute that is linked to the item (Feinberg, [0015]: "the relationship recommendation application configures to identify subgraphs isomorphic to a piece of motif data by generating isomorphic node data based on the attributes of the nodes within the subgraph, generating isomorphic edge data based on the attributes of the edge within the subgraph, generating isomorphic motif node data based on the attributes of the nodes within the piece of motif data, generating isomorphic motif edge data based on the attributes of the edges within the piece of motif data, and generating a similarity score for the subgraphs relative to the piece of motif data based on the isomorphic node data, the isomorphic edge data, the isomorphic motif node data, and the isomorphic motif edge data").
Regarding Claim 6, the rejection of Claim 5 is incorporated. The Feinberg/Canedo/Ferguson combination teaches:
wherein the respective entity of the entity group corresponds to the researcher and the item corresponds to the research journal with which the researcher is associated (Feinberg, [0036]: "Networks of collaborators are also becoming increasingly important in the field of biomedicine, as scientists have increasingly mapped protein and gene interaction networks, forming the interactome for various species. The academic publications and products created by these teams typically have multiple authors and other collaborators. ... [F]or a graph database representing a research environment, the nodes can represent individual researchers while the edges represent collaboration ( e.g. co-authorship) between researchers," where Feinberg's individual researcher and academic publication correspond to the instant entity and item), and
the set of intermediate nodes include a set of key-phrase nodes and a set of topic nodes (Feinberg, [0038]: "Concept description data can be provided that includes keywords and concepts of a particular problem to be solved by identifying a subgraph of the graph database that fulfills the problem. ... [T]he keywords and concepts are mapped to the attributes describing the nodes and edges within the graph database in order to identify the relevant subgraphs," where Feinberg's keyword-attributed and concept-attributed nodes correspond to the instant key-phrase and topic nodes, respectively, where concept attributes support further modeling, as in: [0038]: "the relationship recommendation system can be employed to combine topic modeling and/or natural language processing for a particular concept description").
Regarding Claim 8, the rejection of Claim 1 is incorporated. The Feinberg/Canedo/Ferguson combination teaches:
wherein the querying the graph database comprises: determining a set of distinct entity nodes in the graph database; determining a set of distinct item nodes in the graph database (Feinberg, [0051]: "clusters representing a certain community or communities can be extracted as a subgraph and this subgraph can be subjected to additional clustering" and [0036]: "for a graph database representing a research environment, the nodes can represent individual researchers while the edges represent collaboration (e.g. co-authorship) between researchers," where Feinberg's nodes representing individual researchers corresponds to the instant entity nodes and where Fienberg's research networks comprise co-authorship edges between individual researchers and their collaborators, as in [0037]: "research networks can spawn directed citation graphs where nodes are either authors or papers," where Feinberg's entities are referred to as sets loaded into the database, as in [0008]: "the source data describes a collaborative environment including a set of entities and a set of relationships describing the relationships between the entities in the set of entities");
performing, for each pair of an entity node of the set of distinct entity nodes and an item node of the set of distinct item nodes, a set of operations (Feinberg, [0054]: "the similarity score can be framed as a discrete optimization program by calculating a similarity score pairwise by node with
n
i
and
n
j
corresponding to nodes in the two subsets
N
1
and
N
2
for the subgraphs") that includes:
a first operation to extract a first subgraph of a plurality of subgraphs, wherein the information in the extracted first subgraph includes: a set of intermediate nodes, and a set of edges, each connecting a pair of intermediate nodes of the set of intermediate nodes (Feinberg, [0051]: "clusters representing a certain community or communities can be extracted as a subgraph and this subgraph can be subjected to additional clustering (such as by sequential clustering techniques, topic modeling, or other distinguishing analysis) to identify higher resolution attributes ( e.g., sub-fields) and/or to identify motifs within that subgraph. These higher resolution attributes can then be incorporated into the nodes to describe the multiple levels of clustering for the nodes"); and
a second operation to trim one or more intermediate nodes from the set of intermediate nodes, based on a relevancy attribute assigned to each edge of the set of edges (Feinberg, [0018]: "the relationship recommendation application configures the processor to rank a subgraph based on the attributes of the set of node data and the attributes of the set of edge data for the set of node data and the set of edge data in the subgraph" and [0076]: "subgraphs can be ranked based on the relevance or productivity of those subgraphs as determined by its attributes in the context of the graph database as a whole" and [0026]: "the method further includes ranking the subgraphs generated based on the search node and the additional edges using the relationship recommendation server system and recommending those relationships corresponding to the additional edges forming subgraphs have a rank exceeding a threshold value"); and
extract the set of subgraphs from the plurality of subgraphs by removing one or more duplicate subgraphs from the plurality of subgraphs (Feinberg, [0066]: "the condition that un occur in
S
i
coupled with the condition that un does not occur in the graph database, the augmented graph database is equivalent to finding occurrences of
S
i
that contain u and n and then removing those occurrences already existing within the graph database").
Claim 18 incorporates substantively all limitations of Claim 8 in non-transitory computer-readable storage media form and is rejected under the same rationale as Claim 8.
Claim 19 incorporates all limitations of Claim 16 without further limitation and is further rejected for being an improper depend claim, as indicated above.
Regarding Claim 12, the rejection of Claim 1 is incorporated. The Feinberg/Canedo/Ferguson combination teaches:
wherein the computation of the label score comprises of: normalizing each of the first score, the second score, the third score, and the fourth score (Feinberg, [0053]: "a similarity score
d
for attribute
a
for the motifs can be defined by various metrics and norms. As one example, one can define the similarity score
d
for attribute a as:
d
a
M
1
,
M
2
=
∑
n
∈
N
1
a
n
-
a
g
n
2
+
∑
e
∈
E
1
a
e
-
a
g
e
2
1
2
and
d
M
1
,
M
2
=
∑
a
c
a
d
a
M
1
,
M
2
2
1
2
where
c
a
is a constant for attribute
a
. In a number of embodiments,
c
a
is a constant utilized to weight attributes related to the nodes and edges within the motifs," where Feinberg's similarity score
d
for attribute
a
corresponds to a normalized score), and
summing each of the normalized first score, the normalized second score, the normalized third score, and the normalized fourth score (Feinberg, [0061]: "the commonly occurring substructures (e.g. motifs) are identified not only with respect to their occurrences within graph databases
{
G
i
,
…
,
G
n
}
and similarity scores but also with a set of functions
{
f
*
G
1
,
…
,
f
*
G
n
}
utilized to select the subgraphs" and [0066]: "A ranking function can then be defined using the identified subgraphs as
r
a
n
k
n
u
=
∑
i
f
*
S
i
u
/
d
S
i
u
,
S
i
," where Feinberg's ranking function reasonably suggests the instant sum of normalized scores).
Regarding Claim 15, the rejection of Claim 1 is incorporated. The Feinberg/Canedo/Ferguson combination teaches:
wherein the one or more items in the recommendation are determined based on an association of the one or more items with the one or more subgraphs (Feinberg, [0036]: "Networks of collaborators are also becoming increasingly important in the field of biomedicine, as scientists have increasingly mapped protein and gene interaction networks, forming the interactome for various species. The academic publications and products created by these teams typically have multiple authors and other collaborators. ... [F]or a graph database representing a research environment, the nodes can represent individual researchers while the edges represent collaboration ( e.g. co-authorship) between researchers," where Feinberg's co-authorship of collaborators corresponds to the instant association).
Claims 13 and 14 are rejected under 35 U.S.C. 103 as being unpatentable over Feinberg, et al. (US 2014/0280224 A1, hereinafter "Feinberg") in view of Martinez Canedo, et al. (US 2019/0095806 A1, hereinafter "Canedo") in view of Ferguson, et al., "An Investigation of Term Weighting Approaches for Microblog Retrieval" (hereinafter "Ferguson") in view of Reinanda, "Entity associations for search" (hereinafter "Reinanda")
Regarding Claim 13, the rejection of Claim 1 is incorporated.
The Feinberg/Canedo/Ferguson combination teaches generating a training dataset to include the set of subgraphs and the label score for each subgraph of the set of subgraphs.
The Feinberg/Canedo/Ferguson combination does not explicitly teach partitioning the training dataset into the respective entity-specific subsets of the training dataset, cross-validating each trained ML regression model of the set of ML regression models on at least one subset of the entity-specific subsets that is not used in the training of the respective ML regression model, and determining a set of adjusted scores associated with the set of subgraphs, based on the cross-validation.
However, Reinanda teaches:
partitioning the training dataset into the respective entity-specific subsets of the training dataset (Reinanda, p. 105, 6.3.1 Data and annotations: "In order to have enough training data for each entity, the collection was split based on per-entity cut-off points in time");
cross-validating each trained ML regression model of the set of ML regression models on at least one subset of the entity-specific subsets that is not used in the training of the respective ML regression model (Reinanda, p. 107, 6.3.2 Experiments, Main experiment: Unseen entities: "We randomly split the query entities into five parts and divide the training data accordingly. For every iteration we train on the training data consisting only of document-entity pairs of the corresponding entity split and test on the remaining split. We perform this procedure five times, resulting in a 5-fold cross-validation," where Reinanda's training involves the gradient boosted decision trees (GBDT) ML algorithm); and
determining a set of adjusted scores associated with the set of subgraphs, based on the cross-validation (Reinanda, p. 106, 6.3.3 Evaluation: "we use the evaluation metrics introduced in the TREC KBA track for the vital filtering task:
F
m
a
c
r
o
, and maximum scaled utility (SU). We also compute precision (P), recall (R), and F measure: the average of the harmonic mean of precision and recall over topics," where Reinanda's metrics and scores correspond to the instant set of adjusted scores).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the prior teachings of the Feinberg/Canedo/Ferguson combination regarding generating a training dataset to include the set of subgraphs and the label score for each subgraph of the set of subgraphs with the teachings of Reinanda regarding partitioning the training dataset into the respective entity-specific subsets of the training dataset, cross-validating each trained ML regression model of the set of ML regression models on at least one subset of the entity-specific subsets that is not used in the training of the respective ML regression model, and determining a set of adjusted scores associated with the set of subgraphs, based on the cross-validation.
The motivation to do so would be to facilitate tuning ML model hyperparameters for performance (Reinanda, p. 107, 6.3.5 Parameters and settings: "Our approach involves two sets of hyperparameters. The first set deals with the machine learning algorithm of our choice. GBDT depends on two key parameters: the number of trees, k, and the maximum depth of each tree, d") and to facilitate evaluating algorithm performance (Reinanda, p. 111, Table 6.6: "Results of cross-validation experiments with unseen entities, in terms of
F
m
a
c
r
o
(top), P (middle), and R (bottom)" and p. 110, 6.4.2 Main experiment: Unseen entities: "Averaged over all folds, our approach also achieves the best performance. The differences between the performance of different methods in the unseen entities setting is very small in terms
F
m
a
c
r
o
. Overall, the learned model tends to be precision-oriented with some loss in recall").
Regarding Claim 14, the rejection of Claim 13 is incorporated. The Feinberg/Canedo/Ferguson/ Reinanda combination teaches:
wherein the one or more subgraphs associated with the second entity are determined by: sorting a first subset of ... scores of the set of ... scores associated with the second entity into a list (Feinberg, [0058]: "the identification stage includes enumerating subgraphs of
G
into a list
S
=
{
S
1
,
…
,
S
n
}
with a function n where
n
S
i
is the number of times subgraph
S
i
occurred in
G
(in several embodiments as an equivalence class)");
updating the list of the first subset of ... scores by indexing the prediction score for the unseen subgraph in the list (Feinberg, [0058]: "
Z
S
i
... gives the statistical significance of
S
i
. This statistical significance can be utilized to rank the importance of
S
i
within the graph database," where the index of Feinberg's
Z
S
i
corresponds to the instant indexing); and
comparing the prediction score in the updated list with each ... score of the first subset of ... scores, wherein the one or more subgraphs are determined based on the comparison (Feinberg, [0017]: "the relationship recommendation application further configures the processor to rank the subgraphs generated based on the search node and the additional edges and recommend those relationships corresponding to the additional edges forming subgraphs have a rank exceeding a threshold value").
Reinanda further teaches:
sorting a first subset of adjusted scores of the set of adjusted scores associated with the second entity into a list (Reinanda, p. 110, 6.4.2 Main experiment: Unseen entities: "The results of our experiments with unseen entities are detailed in Table 6.6. Our approach performs best on almost all folds in terms of
F
m
a
c
r
o
, gaining significant improvements compared to other approaches on Fold1 and Fold3," where Reinanda's Table 6.6 displays the sorted folds of validation data);
updating the list of the first subset of adjusted scores by indexing the prediction score for the unseen subgraph in the list (Reinanda, p. 107, 6.3.3 Evaluation: "The main evaluation metric,
F
m
a
c
r
o
, is defined as the maximum of the harmonic mean of averaged precision and recall computed at every possible threshold θ which separates vital and non-vital documents," where Reinanda's computing at possible thresholds corresponds to the instant updating);
comparing the prediction score in the updated list with each adjusted score of the first subset of adjusted scores, wherein the one or more subgraphs are determined based on the comparison (Reinanda, p. 108, 6.4.1 Main experiment: Long-tail entities: "We compute the best threshold for each approach, determine its per-entity performance using this cutoff, and then aggregate the performance by averaging the per-entity scores," where Reinanda's computing the best threshold corresponds to the instant comparing the prediction score).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the teachings of the Feinberg/Reinanda combination regarding determining adjusted scores for subgraphs based on cross-validation with the further teachings of Reinanda regarding sorting and comparing the adjusted scores and prediction score.
The motivation to do so would be to ensure that model evaluation comparison are fair by being based on confidence thresholds providing a maximum score (Reinanda, p. 106, 6.3.3 Evaluation: "The main evaluation metric,
F
m
a
c
r
o
, is defined as the maximum of the harmonic mean of averaged precision and recall computed at every possible threshold θ which separates vital and non-vital documents: max(avg(P), avg(R)). The motivation behind this is evaluation setup is as follows. A filtering system will have a single confidence threshold θ for which the classification performance is maximized. Different systems might have different optimal confidence score calibrations, hence choosing the maximum scores with respect to each system’s best threshold would ensure the fairest comparison").
Claim 7 is rejected under 35 U.S.C. 103 as being unpatentable over Feinberg, et al. (US 2014/0280224 A1, hereinafter "Feinberg") in view of Martinez Canedo, et al. (US 2019/0095806 A1, hereinafter "Canedo") in view of Ferguson, et al., "An Investigation of Term Weighting Approaches for Microblog Retrieval" (hereinafter "Ferguson") in further view of Schaefermeier, et al., "Topic space trajectories (hereinafter "Schaefermeier").
Regarding Claim 7, the rejection of Claim 1 is incorporated. The Feinberg/Canedo/Ferguson combination teaches:
wherein a relevancy weight is assigned to ... each key-phrase node of the set of key-phrase nodes and a ... topic node of the set of topic nodes (Feinberg, [0053]: "one can define the similarity score d for attribute a as: ... where
c
a
is a constant for attribute a. In a number of embodiments,
c
a
is a constant utilized to weight attributes related to the nodes and edges" where Feinberg's keywords and concepts are used for determining relevant subgraphs, as in: [0038]: "the keywords and concepts are mapped to the attributes describing the nodes and edges within the graph database in order to identify the relevant subgraphs" and [0039]: "the similarity score is based on one or more attributes identified as relevant to a particular function (i.e. concept description data) that is the subject of the relationship recommendation process. ... In this example, the most relevant attribute could represent field of study, and with the difference between the fields of study being the similarity score in the identification of motifs").
The Feinberg/Canedo/Ferguson combination teaches weighted keyword and topic attributes of entity nodes used in calculating entity similarity scores.
The Feinberg/Canedo/Ferguson combination does not explicitly teach a relevancy weight is assigned to an edge between each key-phrase node of the set of key-phrase nodes and a corresponding topic node of the set of topic nodes.
However, Schaefermeier teaches:
wherein a relevancy weight is assigned to an edge between each key-phrase node of the set of key-phrase nodes and a corresponding topic node of the set of topic nodes (Schaefermeier, p. 5762: "We employ non-negative matrix factorization (NMF) ... for discovering paper topics and embedding papers into the topic space
R
t
for a chosen dimensionality
t
∈
N
>
0
. We start with a word-document matrix
V
∈
R
≥
0
n
×
d
with
n
being the number of words and d the number of documents. In this matrix, each column is a document vector in which each component is a word weight. A word weight can be, for example, the term frequency in the document or the product of the term frequency and inverted document frequency (tf-idf). NMF finds an approximate factorization
V
≈
W
H
. ... After the minimization the left factor
W
∈
R
≥
0
n
×
t
contains in each column a topic vector of word weights," where Schaefermeier's word corresponds to the instant key-phrase in later handling by way of ranked terms, as in p. 5771: "Table 1 Topics identified through NMF. Each topic is represented here by the top ranked terms").
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the teachings of the Feinberg/ Reinanda combination regarding a relevancy weight being assigned to each key-phrase node of the set of key-phrase nodes and a topic node of the set of topic nodes with those of Schaefermeier regarding a relevancy weight being assigned to an edge between each key-phrase node of the set of key-phrase nodes and a corresponding topic node of the set of topic nodes.
The motivation to do so would be to facilitate identifying relevant topics using an interpretable and semi-automated technique (Schaefermeier, p. 5760, Introduction: "we employ non-negative matrix factorization (NMF), a model whose topics are comparatively well comprehensible.... Moreover, we semi-automatically select an appropriate number of topics based on a coherence measure" and p. 5770, Interpreting topic space trajectories, Found topics: "Most topics found through NMF are well interpretable and often clearly correspond to one specific research area").
Claim 11 is rejected under 35 U.S.C. 103 as being unpatentable over Feinberg, et al. (US 2014/0280224 A1, hereinafter "Feinberg") in view of Martinez Canedo, et al. (US 2019/0095806 A1, hereinafter "Canedo") in view of Ferguson, et al., "An Investigation of Term Weighting Approaches for Microblog Retrieval" (hereinafter "Ferguson") in further view of Marvit, et al. (US 2009/0094233 A1, hereinafter "Marvit").
Regarding Claim 11, the rejection of Claim 1 is incorporated. The Feinberg/Canedo/Ferguson combination teaches:
a set of k-scores for a set of key-phrase nodes of one or more first subgraphs of the set of subgraphs ..., the one or more first subgraphs being associated with the third entity (Feinberg, [0038]: "the relationship recommendation system can be employed ... to determine a 'team' of nodes ... that will be most likely to form a productive solution to the concept description. Taking the previous examples, this team of nodes can represent a team of researchers," where Feinberg's most likely team or researchers associated with keywords corresponds to the instant relative key-phrase score, and where Feinberg's team corresponds to the instant third entity).
The Feinberg/Canedo/Ferguson combination does not teach further comprising determining the fourth score by: sorting a set of k-scores for a set of key-phrase nodes of one or more first subgraphs of the set of subgraphs to select a set of top k-scores, the one or more first subgraphs being associated with the third entity; determining a sum of the set of top k-scores and a sum of the set of k- scores; and determining a second ratio of the sum of the set of top k-scores to the sum of the set of k-scores to determine the fourth score.
However, Marvit teaches:
further comprising determining the fourth score by: sorting a set of k-scores for a set of key-phrase nodes ... to select a set of top k-scores ...; determining a sum of the set of top k-scores and a sum of the set of k- scores (Marvit, [0079]: "clustering engine 210, identifies clusters by sorting words of a set of words into clusters. In one example, the words
{
w
i
}
of set
W
are sorted according to affinities
*
Aff
w
i
,
w
j
, such as differential or directional affinities. In another example, the words { w,} are sorted according to an aggregation function, such as the sum, of affinities of word w, to each member of a distinct set of words Q," where Marvit's words are keywords, as in: [0016]: Words of a document are selected as keywords of the document. The documents are clustered according to the keywords"); and
determining a second ratio of the sum of the set of top k-scores to the sum of the set of k-scores to determine the fourth score (Marvit, [0054]: "word recommender 48 receives a seed word and identifies words that have an affinity with the seed word that is greater than a threshold affinity. The threshold affinity may have any suitable value, such as greater than or equal to 0.25, 0.5, 0.75, or 0.95," where Marvit's affinity is defined as a co-occurrence ratio, as in [0026]: "The basic affinity (A) between terms A and B may be defined as the ratio of the number of co-occurrence contexts that include both terms A and B over the number of co-occurrence contexts that include either of the terms A or B").
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the teachings of Feinberg/Canedo/Ferguson combination regarding a set of k-scores for a set of key-phrase nodes of one or more first subgraphs of the set of subgraphs with those of Marvit regarding determining the fourth score by sorting a set of k-scores to select a set of top k-scores, determining a sum of the set of top k-scores and a sum of the set of k- scores, and determining a second ratio therefrom.
The motivation to do so would be to facilitate studying relationships between people within social networks using an affinity score to filter uninformative names (Marvit, [0090]: "clustering analyzer 214 may use affinity clustering to study a social network. ... Examples of such pages include correspondence (such as letters, emails, and instant messages), memos, articles, and meeting minutes. These pages 50 may include words comprising user identifiers (such as names) of people of a social network. Clusters of names may be identified to analyze relationships among the people of the network. In one example, differential affinity clustering may be used to filter out names that appear most pages 50 without providing information, such as names of system administrators").
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure:
Alsentzer, et al., "Subgraph Neural Networks," teach a subgraph neural network technique, SUBGNN, that propagates neural messages between a subgraph's components and sampled patches from the underlying graph to yield subgraph representations for use in subgraph prediction tasks such as user profiling.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to ROBERT N DAY whose telephone number is (703)756-1519. The examiner can normally be reached M-F 9-5.
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, Kakali Chaki can be reached at (571) 272-3719. 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.
/R.N.D./Examiner, Art Unit 2122
/KAKALI CHAKI/Supervisory Patent Examiner, Art Unit 2122
