DETAILED ACTION
Claims 1-20 are pending in the present application and are under examination on the merits. This communication is the first action on the merits (FAOM).
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 .
Information Disclosure Statement
Applicant filed an Information Disclosure Statement (IDS) on 6/12/2025. This filing is in compliance with 37 C.F.R. 1.97.
As required by M.P.E.P. 609(C), the applicant's submission of the Information Disclosure Statement is acknowledged by the examiner and the cited references have been considered in the examination of the claims now pending. As required by M.P.E.P. 609(C), a copy of the PTOL -1449 form, initialed and dated by the examiner, is attached to the instant office action.
Claim Rejections - 35 USC§ 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to a judicial exception (i.e., a law of nature, a natural phenomenon, or an abstract idea) without significantly more. Here, under considerations of the broadest reasonable interpretation of the claimed invention, Examiner finds that the Applicant invented a method and system for generating a knowledge graph based on survey response data. Examiner formulates an abstract idea analysis, following the framework described in the MPEP as follows:
Step 1: The claims are directed to a statutory category, namely a "method" (claims 17-20) and "system" (claims 1-16).
Step 2A - Prong 1: The claims are found to recite limitations that set forth the abstract idea(s), namely, regarding claim 1:
generate a knowledge graph comprising a plurality of nodes corresponding to a predefined ontology of topics and a plurality of edges indicating relationships between the plurality of nodes, wherein the predefined ontology is configured to normalize topics across different entities utilizing person nodes representing survey respondents, activity nodes representing respondent interactions, experience nodes representing survey question responses, and operational nodes representing survey data attributes;
receive, from one or more client devices, survey data comprising question data and response data for one or more electronic survey questions;
extract topics from the survey data and determine connections between the extracted topics and the plurality of nodes in the knowledge graph by linking the extracted topics to the person nodes, the activity nodes, the experience nodes, or the operational nodes in the knowledge graph; and
generate a digital benchmark for a first set of data relative to a second set of data according to the connections between the extracted topics in the knowledge graph.
Independent claims 11 and 17 recites substantially similar claim language.
Dependent claims 2-10, 12-16, and 18-20 recite the same or similar abstract idea(s) as independent claims 1, 11, and 17 with merely a further narrowing of the abstract idea(s) to particular data characterization and/or additional data analyses performed as part of the abstract idea.
The limitations in claims 1-20 above falling well-within the groupings of subject matter identified by the courts as being abstract concepts, specifically the claims are found to correspond to the category of:
"Certain methods of organizing human activity- fundamental economic principles or practices (including hedging, insurance, mitigating risk); commercial or legal interactions (including agreements in the form of contracts; legal obligations; advertising, marketing or sales activities or behaviors; business relations); managing personal behavior or relationships or interactions between people (including social activities, teaching, and following rules or instructions)" as the limitations identified above are directed to generating a knowledge graph based on survey response data and thus is a method of organizing human activity including at least commercial or business interactions or relations and/or a management of user personal behavior; and/or
"Mental processes - concepts performed in the human mind (including an observation, evaluation, judgement, opinion)" as the limitations identified above include mere data observations, evaluations, judgements, and/or opinions, e.g. including user observation and generating a knowledge graph based on survey response data, which is capable of being performed mentally and/or using pen and paper.
Step 2A - Prong 2: Claims 1-20 are found to clearly be directed to the abstract idea identified above because the claims, as a whole, fail to integrate the claimed judicial exception into a practical application, specifically the claims recite the additional elements of:
" A non-transitory computer readable storage medium comprising instructions that, when executed by at least one processor, cause a computing device to: / A system comprising: at least one processor; and at least one non-transitory computer readable storage medium comprising instructions that, when executed by at least one processor, cause the system to: / A computer-implemented method comprising " (claims 1, 11, and 17) however the aforementioned elements merely amount to generic components of a general purpose computer used to "apply" the abstract idea (MPEP 2106.0S(f)) and thus fails to integrate the recited abstract idea into a practical application, furthermore the high-level recitation of receiving data from a generic "computer" is at most an attempt to limit the abstract to a particular field of use (MPEP 2106.0S(h), e.g.: "For instance, a data gathering step that is limited to a particular data source (such as the Internet) or a particular type of data (such as power grid data or XML tags) could be considered to be both insignificant extra-solution activity and a field of use limitation. See, e.g., Ultramercial, 772 F.3d at 716, 112 USPQ2d at 1755 (limiting use of abstract idea to the Internet); Electric Power, 830 F.3d at 1354, 119 USPQ2d at 1742 (limiting application of abstract idea to power grid data); Intellectual Ventures I LLC v. Erie lndem. Co., 850 F.3d 1315, 1328-29, 121 USPQ2d 1928, 1939 (Fed. Cir. 2017) (limiting use of abstract idea to use with XML tags).") and/or merely insignificant extra-solution activity (MPE 2106.05(g)) and thus further fails to integrate the abstract idea into a practical application;
Step 2B: Claims 1-20 do not include additional elements that are sufficient to amount to significantly more than the judicial exception because the additional elements as described above with respect to Step 2A Prong 2 merely amount to a general purpose computer that attempts to apply the abstract idea in a technological environment (MPEP 2106.0S(f)), including merely limiting the abstract idea to a particular field of use of analysis of a knowledge graph via a " computer readable storage medium" and “processor,” as explained above, and/or performs insignificant extra-solution activity, e.g. data gathering or output, (MPEP 2106.0S(g)), as identified above, which is further found under step 2B to be merely well-understood, routine, and conventional activities as evidenced by MPEP 2106.0S(d)(II) (describing conventional activities that include transmitting and receiving data over a network, electronic recordkeeping, storing and retrieving information from memory, electronically scanning or extracting data from a physical document, and a web browser's back and forward button functionality). Therefore, similarly the combination and arrangement of the above identified additional elements when analyzed under Step 2B also fails to necessitate a conclusion that the claims amount to significantly more than the abstract idea directed to generating a knowledge graph based on survey response data.
Claims 1-20 are accordingly rejected under 35 USC§ 101 because the claimed invention is directed to a judicial exception (i.e., a law of nature, a natural phenomenon, or an abstract idea(s)) without significantly more.
Note: The analysis above applies to all statutory categories of invention. As such, the presentment of any claim otherwise styled as a machine or manufacture, for example, would be subject to the same analysis.
For further authority and guidance, see:
MPEP § 2106
https://www.uspto.gov/patents/laws/examination-policy/subject-matter-eligibility
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 of this title, 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.
Claims 1-20 are rejected under 35 U.S.C. 103 as being unpatentable over U.S. Patent Number 11557276 to Bender et al. (hereafter referred to as Bender) in view of U.S. Patent Application Publication Number 2022/0384001 to Gnanasambandam et al. (hereafter referred to as Gnanasambandam) and in further view of U.S. Patent Application Publication Number 2021/0224346 to Peng et al. (hereafter referred to as Peng).
As per claim 1, Bender teaches:
A non-transitory computer readable storage medium comprising instructions that, when executed by at least one processor, cause a computing device to: (claim 9 teaches a tangible, non-transitory, machine-readable medium storing instructions that, when executed by one or more processors, effectuate operations comprising: obtaining, with a computer system, a set of ontology graphs comprising a first ontology graph associated with a first domain category value and a second ontology graph associated with a second domain category value).
generate a knowledge graph comprising a plurality of nodes corresponding to a predefined ontology of topics (Col. 11 line 60 - col. 12 line 11 teaches an indicated expertise score or other score associated with a domain to determine a hierarchy or other order between different ontologies or knowledge graphs. A set of ontology graphs organized in a hierarchy may be used as part of an index for a knowledge fabric, which may include a set of documents or other data, the ontology system(s) and indices used to organize the set of documents or other data, or the functions used use the ontology system(s) or indices used to retrieve information from the set of documents or other data. By using a knowledge fabric that is organized by a set of ontology graphs, some embodiments may quickly navigate through different knowledge domains or different classes within those domains to retrieve relevant queries for a specific user).
and a plurality of edges indicating relationships between the plurality of nodes, (Col. 9 lines 20-43 teach use of various types of data to generate or otherwise update the ontology data stored in the ontology data repository 230. The data may include a set of existing ontology data 211, a set of natural-language text documents 212, or a set of structured data 214. The existing knowledge graph may be stored in various ways, such as in a relational data structure, and may be imported into the ontology data repository 230. Different data types may be combined to update an ontology data model, such as one stored in an ontology data model record 231. The ontology data model record may store values for record fields such as object categories, relationships between the categories, directional indicators of the relationships, or the like. Alternatively, or in addition, the ontology data model may be stored in a knowledge graph such as a knowledge graph 232, which may be formatted in a specified ontology data model. (Examiner asserts that the relational data structure that combines data types including object categories, relationships, etc., is equivalent to an experience node based upon Paragraph Number [0058] of Applicant's specification which provides: "the survey knowledge graph system 102 determines respondent feedback such as ratings, opinions, sentiment, emotion, or other responses that indicate experiences of users with activities or topics")).
wherein the predefined ontology is configured to normalize topics across different entities (Col. 61 lines 1-32 teach some embodiments may perform self-attention operations by computing a set of key vectors, query vectors, or value vectors determined using a set of embedding vectors and positional encoding vectors. In some embodiments, the key, query, and value vectors may be determined during a training operation of a transformer model. After training, some embodiments may compute attention values using a function that takes, as input(s), the sets of key, query, and value vectors. For example, using an attention-determining function may include computing a product of a first element of a query vector with a second element of a key vector, where the product may be computed as part of a dot product determination between a query vector of a first n-gram with key vectors of other n-grams of a sequence of n-grams, where the output of the dot product may be further processed to determine an attention value. Various modifications to the output vector(s) may be performed, such as determining a root of the output, performing a normalization of the root by performing a set of softmax operations, or the like).
and operational nodes representing survey data attributes (Col. 4 lines 31-57 teach such other information can include the document's origin, metadata associated with the document, a data format, or the like. A domain vector for a document can indicate various types of information relevant to the usefulness of the document, such as an associated expertise level for each of a plurality of domains, a count of words, a count of words having more than a specified number of syllables, or the like. Col. 7 line 62 - Col. 8 line 10 teaches store ontology data in a set of SQL tables of the ontology database 138, where each record of the SQL table may represent a vertex record and include, as table fields, parent vertex identifiers, child vertex identifiers, categories indicating relationship types between vertices, scores associated with the relationship category, or the like. Some embodiments may store data in a combination of relational and non-relational databases. Col. 17 lines 55-65 teach categorize a vertex group determined from a clustering method as a first type of vertex group, where vertices of a vertex group of the first type of vertex group may be associated with vectors categorized as being part of a same cluster. In some embodiments, the vertex group may represent a ‘concept’ in a domain, where the concept may be shared amongst multiple classes of the domain).
extract topics from the survey data and determine connections between the extracted topics and the plurality of nodes in the knowledge graph by linking the extracted topics to the person nodes, the activity nodes, the experience nodes, or the operational nodes in the knowledge graph (Col. 18 lines 26-43 teach some embodiments may determine a hierarchy of ontologies based on one or more vertex groups associated with a vector cluster via an edge connection between the vertex group and one or more shared connections. Some embodiments may determine that a first vertex group of a first ontology graph may be associated with a vertex of a second ontology graph via one or more shared vertices. For example, a first vertex group may include a set of vertices corresponding with a first set of learned representations that includes the embedding vector [x.sub.1, x.sub.2, x.sub.3]. Some embodiments may determine that the embedding vector corresponds with a vertex in a second ontology graph, and, in response, determine that the first vertex group is associated with the vertex in the second ontology graph. Additionally, if the vertex in the second ontology graph is part of a second vertex group, some embodiments may determine that the first vertex group is associated with the second vertex group of the ontology graph. (See also Col. 4 lines 31-57 and Col. 7 line 62 - Col. 8 line 10)).
generate a digital benchmark for a first set of data relative to a second set of data according to the connections between the extracted topics in the knowledge graph (Col. 6 lines 6-33 teach an ontology data model or knowledge graph organized by the ontology data model may improve the relevance of retrieved documents by accounting for a user's domain expertise or specific interests. Such operations may be especially useful in specialized applications where similar concepts may be disclosed in documents at differing levels of domain expertise, differing levels of security classification, or with differing amounts of relevance to subdomains. For example, if a user associated with a first hierarchical expertise level performs a search, some embodiments may obtain a first document associated with the first hierarchical expertise level and a second document associated with a second hierarchical expertise level. Some embodiments may then provide a user with the document associated with the first hierarchical expertise level. Additionally, by encoding relative levels of domain expertise or other domain-specific relationships in graph edges that indicate cross-domain relations, some embodiments may improve the speed and accuracy of responses to queries for information or provide other aspects of expert guidance. It should be emphasized, though, that not all embodiments necessarily provide these advantages, as there are several independently useful ideas described herein, and some implementations may only apply a subset of these techniques. As used in this disclosure, the term “ontology” may be used interchangeably with the term “ontology graph,” unless otherwise indicated, where an entry of an ontology may include a vertex of the ontology).
Bender teaches receiving requests for data and utilizing knowledge graphs to gather, store, and profile data but does not explicitly teach extracting data through survey questions and real-time communications as described by the following citations from Gnanasambandam:
activity nodes representing respondent interactions (Paragraph Number [0287] teaches receiving a first data including user registration data (block 702); and providing a health assessment and receiving second data including health assessment answers (block 704). In various embodiments, the health assessment is a micro-survey with dynamically formulated questions presented to the user. Paragraph Number [0303] teaches the cognitive agent 110 can answer a library of questions and provide content for many questions a user has as it related to diabetes. The information provided for purposes of educating a user is based on an overall health plan of the user, which is based on meta data analysis of interactions with the user, and an analysis of the education level of the user. Paragraph Number [0304] teaches the user inputs text using the input box 916, which instructs the user to “Talk to me or type your question”).
experience nodes representing survey question responses (Paragraph Number [0287] teaches receiving a first data including user registration data (block 702); and providing a health assessment and receiving second data including health assessment answers (block 704). In various embodiments, the health assessment is a micro-survey with dynamically formulated questions presented to the user. Paragraph Number [0303] teaches the cognitive agent 110 can answer a library of questions and provide content for many questions a user has as it related to diabetes. The information provided for purposes of educating a user is based on an overall health plan of the user, which is based on meta data analysis of interactions with the user, and an analysis of the education level of the user. Paragraph Number [0304] teaches the user inputs text using the input box 916, which instructs the user to “Talk to me or type your question”).
receive, from one or more client devices, survey data comprising question data and response data for one or more electronic survey questions (Paragraph Number [0287] teaches receiving a first data including user registration data (block 702); and providing a health assessment and receiving second data including health assessment answers (block 704). In various embodiments, the health assessment is a micro-survey with dynamically formulated questions presented to the user. Paragraph Number [0303] teaches the cognitive agent 110 can answer a library of questions and provide content for many questions a user has as it related to diabetes. The information provided for purposes of educating a user is based on an overall health plan of the user, which is based on meta data analysis of interactions with the user, and an analysis of the education level of the user. Paragraph Number [0304] teaches the user inputs text using the input box 916, which instructs the user to “Talk to me or type your question”).
Both Bender and Gnanasambandam are directed to data extraction and semantic analysis. Bender discloses receiving requests for data and utilizing knowledge graphs to gather, store, and profile data. Gnanasambandam improves upon Bender by disclosing extracting data through survey questions and real-time communications. One of ordinary skill in the art would be motivated to further include extracting data through survey questions and real-time communications, to efficiently provide real-time updated information to the knowledge graph. Accordingly, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the system and method of receiving requests for data and utilizing knowledge graphs to gather, store, and profile data in Bender to further utilize extracting data through survey questions and real-time communications as disclosed in Gnanasambandam, since the claimed invention is merely a combination of old elements, and in combination each element merely would have performed the same function as it did separately, and one of ordinary skill in the art would have recognized that the results of the combination were predictable.
Bender teaches receiving requests for data and utilizing knowledge graphs to gather, store, and profile data but does not explicitly teach generating nodes that include person nodes that represent information about a particular user and containing information about that specific user as described by the following citations from Peng:
utilizing person nodes representing survey respondents (Paragraph Number [0076] teaches a user node 602 may correspond to a user of the social-networking system 160 or the assistant system 160. As an example, and not by way of limitation, a user may be an individual (human user), an entity (e.g., an enterprise, business, or third-party application), or a group (e.g., of individuals or entities) that interacts or communicates with or over the social-networking system 160 or the assistant system 160. In particular embodiments, when a user registers for an account with the social-networking system 160, the social-networking system 160 may create a user node 602 corresponding to the user, and store the user node 602 in one or more data stores. Paragraph Number [0079] A user viewing the third-party web interface may perform an action by selecting one of the icons (e.g., “check-in”), causing a client system 130 to send to the social-networking system 160 a message indicating the user's action. In response to the message, the social-networking system 160 may create an edge (e.g., a check-in-type edge) between a user node 602 corresponding to the user and a concept node 604 corresponding to the third-party web interface or resource and store edge 606 in one or more data stores. Paragraph Number [0081] teaches although this disclosure describes edges between a user node 602 and a concept node 604 representing a single relationship, this disclosure contemplates edges between a user node 602 and a concept node 604 representing one or more relationships. As an example and not by way of limitation, an edge 606 may represent both that a user likes and has used at a particular concept. Alternatively, another edge 606 may represent each type of relationship (or multiples of a single relationship) between a user node 602 and a concept node 604 (as illustrated in FIG. 6 between user node 602 for user “E” and concept node 604)).
Both the combination of Bender and Gnanasambandam and Peng are directed to data extraction and semantic analysis. The combination of Bender and Gnanasambandam discloses receiving requests for data and utilizing knowledge graphs to gather, store, and profile data. Peng improves upon the combination of Bender and Gnanasambandam by disclosing generating nodes that include person nodes that represent information about a particular user and containing information about that specific user. One of ordinary skill in the art would be motivated to further include generating nodes that include person nodes that represent information about a particular user and containing information about that specific user, to efficiently connect people and their associated concepts to other nodes in a knowledge graph. Accordingly, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the system and method of receiving requests for data and utilizing knowledge graphs to gather, store, and profile data in the combination of Bender and Gnanasambandam to further utilize generating nodes that include person nodes that represent information about a particular user and containing information about that specific user as disclosed in Peng, since the claimed invention is merely a combination of old elements, and in combination each element merely would have performed the same function as it did separately, and one of ordinary skill in the art would have recognized that the results of the combination were predictable.
As per claim 11, Bender teaches:
A system comprising: at least one processor; and at least one non-transitory computer readable storage medium comprising instructions that, when executed by at least one processor, cause the system to: (claim 9 teaches a tangible, non-transitory, machine-readable medium storing instructions that, when executed by one or more processors, effectuate operations comprising: obtaining, with a computer system, a set of ontology graphs comprising a first ontology graph associated with a first domain category value and a second ontology graph associated with a second domain category value).
The remaining claim limitations are substantially similar to those found in claim 1 and are rejected for the same reasons put forth in regard to claim 1.
As per claim 17, claim 17 recites a computerized method for performing substantially the same steps as presented in regard to claim 1 and is rejected for the same reasons put forth in regard to claim 1.
As per claim 2, the combination of Bender, Gnanasambandam, and Peng teaches each of the limitations of claim 1.
In addition, Bender teaches:
comparing the first set of data to the second set of data within the knowledge graph to identify one or more patterns common to the first set of data and the second set of data (Col. 6 lines 6-33 teach an ontology data model or knowledge graph organized by the ontology data model may improve the relevance of retrieved documents by accounting for a user's domain expertise or specific interests. Such operations may be especially useful in specialized applications where similar concepts may be disclosed in documents at differing levels of domain expertise, differing levels of security classification, or with differing amounts of relevance to subdomains. For example, if a user associated with a first hierarchical expertise level performs a search, some embodiments may obtain a first document associated with the first hierarchical expertise level and a second document associated with a second hierarchical expertise level. Some embodiments may then provide a user with the document associated with the first hierarchical expertise level. Additionally, by encoding relative levels of domain expertise or other domain-specific relationships in graph edges that indicate cross-domain relations, some embodiments may improve the speed and accuracy of responses to queries for information or provide other aspects of expert guidance. It should be emphasized, though, that not all embodiments necessarily provide these advantages, as there are several independently useful ideas described herein, and some implementations may only apply a subset of these techniques. As used in this disclosure, the term “ontology” may be used interchangeably with the term “ontology graph,” unless otherwise indicated, where an entry of an ontology may include a vertex of the ontology).
Bender teaches receiving requests for data and utilizing knowledge graphs to gather, store, and profile data but does not explicitly teach extracting data through survey questions and real-time communications as described by the following citations from Gnanasambandam:
determining one or more anomalies in the survey data based on the one or more patterns ([0300] FIG. 8C illustrates aspects of an additional tool—e.g., a microsurvey—provided to the user that helps gather additional information about the user (e.g., available data). In various embodiments, a micro-survey represent a short targeted survey, where the questions presented in the survey are limited to a respective micro-theory. A microsurvey can be created by the cognitive intelligence platform 102 for several different purposes, including: completing a user profile, and informing a missing parameter during the process of answering an originating question).
A person of ordinary skill would have been motivated to combine these references for the same reasons put forth in regard to claim 1.
As per claim 3, the combination of Bender, Gnanasambandam, and Peng teaches each of the limitations of claim 1.
In addition, Bender teaches:
generate a modification to the predefined ontology of topics based on receiving, from an additional client device, additional survey data comprising additional question data and additional response data for one or more additional electronic survey questions (Col. 21 lines 47-64 teach various types of indices may be constructed or updated based on to an ontology graph, such as an index having a self-balancing tree data structure (“B-tree index”), where a b-tree index may include a set of index nodes starting at root index node. A B-tree index may have an order value m, where each index node has at most m child index nodes, each non-leaf index node has at least m/2 index nodes, and a non-leaf index node having k child index nodes will contain a proportional number of keys for their child index nodes. An index node of a B-tree may include a key value and a pointer to another index node. In some embodiments, the key value of the index node may correspond to one of a pair of n-grams, where a child index node of the index node acting as a leaf index node may include a pointer to or other identifier of the other n-gram of the pair of n-grams. Col. 76 lines 43-65 teach various operations may be performed to retrieve related n-grams of an initial set of n-grams using an index. Some embodiments may search through a self-balancing search tree based on a key, where the key may be an n-gram or a learned representation of the n-gram. Some embodiments may search through the self-balancing search tree by starting at a root of the self-balancing search tree and recursively traversing tree nodes using the key to retrieve a second n-gram or corresponding embedding vector at a leaf node of the self-balancing search tree).
update the knowledge graph by updating one or more of the person nodes, the activity nodes, the experience nodes, or the operational nodes to according to the modification to the predefined ontology of topics (Col. 9 lines 20-43 teach use of various types of data to generate or otherwise update the ontology data stored in the ontology data repository 230. The data may include a set of existing ontology data 211, a set of natural-language text documents 212, or a set of structured data 214. The existing knowledge graph may be stored in various ways, such as in a relational data structure, and may be imported into the ontology data repository 230. Different data types may be combined to update an ontology data model, such as one stored in an ontology data model record 231. The ontology data model record may store values for record fields such as object categories, relationships between the categories, directional indicators of the relationships, or the like. Alternatively, or in addition, the ontology data model may be stored in a knowledge graph such as a knowledge graph 232, which may be formatted in a specified ontology data model. (Examiner asserts that the relational data structure that combines data types including object categories, relationships, etc., is equivalent to an experience node based upon Paragraph Number [0058] of Applicant's specification which provides: "the survey knowledge graph system 102 determines respondent feedback such as ratings, opinions, sentiment, emotion, or other responses that indicate experiences of users with activities or topics")).
As per claim 4, the combination of Bender, Gnanasambandam, and Peng teaches each of the limitations of claim 1.
In addition, Bender teaches:
wherein generating the knowledge graph further comprises generating edge weights indicating dependencies between the plurality of nodes using a graph neural network trained on historical survey response data. (Col. 13 lines 51-65 teaches determining an embedding vector associated with an n-gram using a model based on both the n-gram itself and the context surrounding the n-gram (e.g., other n-grams, syntax, semantics). For example, some embodiments may use neural networks models trained on a set of text of a corpus or other training data to predict n-grams based on other n-grams in a system via a set of attention values for the n-grams, where the attention values may be used to weigh or otherwise modify an output of a neural network. Various models may use different types of neural network models, perform different pre-processing operations, use different operations to determine attention values, or the like. Col. 25 line 62 - Col. 26 line 10 teaches transfer parameters of machine learning model, where the parameters may include a set of neural network parameters such as weights, biases, activation function parameters, or other values of the neurons of a neural network. Once transferred, these parameters may be used by a new instance of the neural network model or other machine learning model. For example, some embodiments can train a BERT-based machine learning model to predict answers based on training queries from a stored library of queries and answers, where the answers for the queries may include semantic search results. Some embodiments may train a machine learning model based on a set of training queries and a corresponding set of training documents that should be retrieved when the system is provided with the set of training queries).
As per claim 5, the combination of Bender, Gnanasambandam, and Peng teaches each of the limitations of claim 1.
In addition, Bender teaches:
determining that the response data comprises a text response to an electronic survey question. (Col. 23 lines 6-19 teach various other associations in the index may be made, such as associating the first vertex with the second document or the third document. By associating documents of a corpus in an index based on hierarchical associations between vertices of knowledge graphs and vertex adjacency in a knowledge graph, some embodiments may increase the speed of document retrieval by using the index. Additionally, as further described below, some embodiments may generate question-answer pairs based on the knowledge graph and include the question-answer pairs in an index. For example, some embodiments may associate a specific query or type of query with a specific document or set of documents. Some embodiments may include this association representing a question-answer pair in the index).
extracting a topic indicated in the text response. (Col. 55 line 64 - Col. 56 line 14 teaches various types of scoring models may be used to score n-grams of a natural-language text document, where n-grams of a natural-language text document may be scored by individually scoring the n-grams or by scoring text sections including the n-grams. In some embodiments, the scoring model may be a model used for extractive summarization methods, where one or more text sections may be selected as summarizing text sections based on the value of the corresponding scores. Some embodiments may determine topic scores for each sentence of a natural language document. In some embodiments, each respective topic score corresponding to a respective text section and may indicate relevance to a topic, where the topic may be determined from a query or an updated query. For example, a first topic of a query may include the phrase “atrial fibulation” based on the query including the phrase “atrial fibulation,” and a second topic of the query may include the acronym “NVAF” based on a set of cross-graph associations between the n-gram “NVAF” and the n-gram “Atrial Fibulation.”).
As per claim 6, the combination of Bender, Gnanasambandam, and Peng teaches each of the limitations of claims 1 and 5.
In addition, Bender teaches:
determine that the topic corresponds to one or more topics in the predefined ontology of topics (Col. 25 line 62 - col. 26 line 25 teaches transfer parameters of machine learning model, where the parameters may include a set of neural network parameters such as weights, biases, activation function parameters, or other values of the neurons of a neural network. Once transferred, these parameters may be used by a new instance of the neural network model or other machine learning model. For example, some embodiments can train a BERT-based machine learning model to predict answers based on training queries from a stored library of queries and answers, where the answers for the queries may include semantic search results. Some embodiments may train a machine learning model based on a set of training queries and a corresponding set of training documents that should be retrieved when the system is provided with the set of training queries. Col. 50 lines 4 - 20 teach the set of task-specific layers 1214 may include a second set of neural network parameters that are updated with a set of training operations to perform various tasks. In some embodiments, the set of training operations may use parameters obtained from a template task library 1204 to update parameters of the set of task-specific layers 1214. The template task library 1204 may include training task parameters such as a training dataset for tasks such as summarization, text classification, language modeling, named entry recognition, text encoding, ontology lookup, natural language generation, question-answering, text representation, or the like. For example, some embodiments may train the learning model 1210 by importing neural network parameters from the template task library 1204 into a set of neural network layers of the set of task-specific layers 1214 and use a first training dataset of the template task library 1204 to train the modified learning model 1210).
associate the topic with a topic node of the plurality of nodes based on the one or more topics in the predefined ontology of topics (Col. 35 lines 24-57 teach after updating a set of ontology graphs by forming cross-graph edges relating vertices of different ontology graphs, some embodiments may provide previously-undetected associations between different vertices. For example, the first cross-graph edge associating the eighth vertex of the third ontology graph with the ninth vertex of the second ontology graph may be used to determine an association between the n-gram “aFib” and the n-gram “Valvular Atrial Fibrillation.” Additionally, some embodiments may detect previously non-established links between vertices of a same ontology graph by using edges associating vertices of ontology graphs having different domains or classes. For example, as shown in the association between the n-gram “Atrial Fibrillation” written in the box 831 and the n-gram “NVAF” written in the box 834, some embodiments may detect an association between the n-gram “atrial fibulation” and the n-gram “NVAF,” where the association may be recorded in one or more of the vertices or otherwise stored in a database (e.g., as an ontological triple). This association may be used to include the n-gram “NVAF” when expanding a query having the n-gram “aFib,” generating a set of expanded queries, or otherwise performing searches to retrieve documents using query that includes the n-gram “aFib” or “Atrial Fibulation.”).
As per claim 7, the combination of Bender, Gnanasambandam, and Peng teaches each of the limitations of claims 1 and 5.
In addition, Bender teaches:
generate, in response to associating the topic indicated in the response data from the one or more electronic survey questions with a topic node, a leaf user node connected by an edge to the topic node (Col. 21 lines 47-64 teach various types of indices may be constructed or updated based on to an ontology graph, such as an index having a self-balancing tree data structure (“B-tree index”), where a b-tree index may include a set of index nodes starting at root index node. A B-tree index may have an order value m, where each index node has at most m child index nodes, each non-leaf index node has at least m/2 index nodes, and a non-leaf index node having k child index nodes will contain a proportional number of keys for their child index nodes. An index node of a B-tree may include a key value and a pointer to another index node. In some embodiments, the key value of the index node may correspond to one of a pair of n-grams, where a child index node of the index node acting as a leaf index node may include a pointer to or other identifier of the other n-gram of the pair of n-grams.).
Bender teaches receiving requests for data and utilizing knowledge graphs to gather, store, and profile data but does not explicitly teach extracting data through survey questions and real-time communications as described by the following citations from Gnanasambandam:
based on respondent data associated with a user of the one or more client devices, the leaf user node comprising a person node type (Paragraph Number [0287] teaches receiving a first data including user registration data (block 702); and providing a health assessment and receiving second data including health assessment answers (block 704). In various embodiments, the health assessment is a micro-survey with dynamically formulated questions presented to the user. Paragraph Number [0303] teaches the cognitive agent 110 can answer a library of questions and provide content for many questions a user has as it related to diabetes. The information provided for purposes of educating a user is based on an overall health plan of the user, which is based on meta data analysis of interactions with the user, and an analysis of the education level of the user. Paragraph Number [0304] teaches the user inputs text using the input box 916, which instructs the user to “Talk to me or type your question”).
A person of ordinary skill would be motivated to combine these references as described in regard to claim 1.
As per claim 8, the combination of Bender, Gnanasambandam, and Peng teaches each of the limitations of claims 1, 5, and 6.
In addition, Bender teaches:
generate an experience node comprising experience data of the response data from the text response and data inferred from the response data from the text response (Col. 9 lines 20-43 teach use of various types of data to generate or otherwise update the ontology data stored in the ontology data repository 230. The data may include a set of existing ontology data 211, a set of natural-language text documents 212, or a set of structured data 214. The existing knowledge graph may be stored in various ways, such as in a relational data structure, and may be imported into the ontology data repository 230. Different data types may be combined to update an ontology data model, such as one stored in an ontology data model record 231. The ontology data model record may store values for record fields such as object categories, relationships between the categories, directional indicators of the relationships, or the like. Alternatively, or in addition, the ontology data model may be stored in a knowledge graph such as a knowledge graph 232, which may be formatted in a specified ontology data model. (Examiner asserts that the relational data structure that combines data types including object categories, relationships, etc., is equivalent to an experience node based upon Paragraph Number [0058] of Applicant's specification which provides: "the survey knowledge graph system 102 determines respondent feedback such as ratings, opinions, sentiment, emotion, or other responses that indicate experiences of users with activities or topics")).
associate the topic indicated in the text response with the topic node by connecting the experience node to the topic node via one or more edges indicating a relationship between the text response and the topic node (Col. 35 lines 24-57 teach after updating a set of ontology graphs by forming cross-graph edges relating vertices of different ontology graphs, some embodiments may provide previously-undetected associations between different vertices. For example, the first cross-graph edge associating the eighth vertex of the third ontology graph with the ninth vertex of the second ontology graph may be used to determine an association between the n-gram “aFib” and the n-gram “Valvular Atrial Fibrillation.” Additionally, some embodiments may detect previously non-established links between vertices of a same ontology graph by using edges associating vertices of ontology graphs having different domains or classes. For example, as shown in the association between the n-gram “Atrial Fibrillation” written in the box 831 and the n-gram “NVAF” written in the box 834, some embodiments may detect an association between the n-gram “atrial fibulation” and the n-gram “NVAF,” where the association may be recorded in one or more of the vertices or otherwise stored in a database (e.g., as an ontological triple). This association may be used to include the n-gram “NVAF” when expanding a query having the n-gram “aFib,” generating a set of expanded queries, or otherwise performing searches to retrieve documents using query that includes the n-gram “aFib” or “Atrial Fibulation.”).
As per claim 9, the combination of Bender, Gnanasambandam, and Peng teaches each of the limitations of claims 1, 5, and 6.
In addition, Bender teaches:
determining that the first set of data corresponds to a first plurality of nodes connected to the topic node, the first plurality of nodes comprising a person node and an experience node (Col. 9 lines 20-43 teach use of various types of data to generate or otherwise update the ontology data stored in the ontology data repository 230. The data may include a set of existing ontology data 211, a set of natural-language text documents 212, or a set of structured data 214. The existing knowledge graph may be stored in various ways, such as in a relational data structure, and may be imported into the ontology data repository 230. Different data types may be combined to update an ontology data model, such as one stored in an ontology data model record 231. The ontology data model record may store values for record fields such as object categories, relationships between the categories, directional indicators of the relationships, or the like. Alternatively, or in addition, the ontology data model may be stored in a knowledge graph such as a knowledge graph 232, which may be formatted in a specified ontology data model. (Examiner asserts that the relational data structure that combines data types including object categories, relationships, etc., is equivalent to an experience node based upon Paragraph Number [0058] of Applicant's specification which provides: "the survey knowledge graph system 102 determines respondent feedback such as ratings, opinions, sentiment, emotion, or other responses that indicate experiences of users with activities or topics"). Col. 21 lines 47-64 teach various types of indices may be constructed or updated based on to an ontology graph, such as an index having a self-balancing tree data structure (“B-tree index”), where a b-tree index may include a set of index nodes starting at root index node. A B-tree index may have an order value m, where each index node has at most m child index nodes, each non-leaf index node has at least m/2 index nodes, and a non-leaf index node having k child index nodes will contain a proportional number of keys for their child index nodes. An index node of a B-tree may include a key value and a pointer to another index node. In some embodiments, the key value of the index node may correspond to one of a pair of n-grams, where a child index node of the index node acting as a leaf index node may include a pointer to or other identifier of the other n-gram of the pair of n-grams.).
determining that the second set of data corresponds to a second plurality of nodes connected to the topic node (Col. 56 line 50-67 teaches as described elsewhere in this disclosure, some embodiments may generate scores for individual n-grams of a document. For example, some embodiments may generate a score for each n-gram of a sentence of a document, and text sections comprising the sentence may be scored based on the individual scores of the n-grams. In some embodiments, the scoring model may include a neural network that determines a sentence score. For example, some embodiments may use a recurrent neural network to determine a learned representation of a sentence with respect to a specific topic, where different RNNs may be used to determine different sentence scores for a same sentence with respect to different topics. Some embodiments may then select a set of text sections that satisfy a score criteria such as a relevance threshold (e.g., in the form of a minimum score threshold) or select a set of text sections based on their rankings to determine which text sections to analyze or display in a user interface (UI). Col. 72 lines 5-17 teach as disclosed elsewhere in this disclosure, some embodiments may access different indices or different portions and index based on a user context parameter, such as one identifying a domain category value. Alternatively, or in addition, some embodiments may apply different scoring systems based on a user context. For example, some embodiments may use a first scoring model to determine scores for a set of sentences of a document, where the scores may indicate a predicted relevance to a first topic. Some embodiments may then use a second scoring model to determine a different set of scores for the same set of sentences of a document, where the second set of scores may indicate a predicted relevance to a second topic).
generating the digital benchmark based on a comparison of first information stored in the person node and the experience node to second information stored in the second plurality of nodes (Col. 6 lines 6-33 teach an ontology data model or knowledge graph organized by the ontology data model may improve the relevance of retrieved documents by accounting for a user's domain expertise or specific interests. Such operations may be especially useful in specialized applications where similar concepts may be disclosed in documents at differing levels of domain expertise, differing levels of security classification, or with differing amounts of relevance to subdomains. For example, if a user associated with a first hierarchical expertise level performs a search, some embodiments may obtain a first document associated with the first hierarchical expertise level and a second document associated with a second hierarchical expertise level. Some embodiments may then provide a user with the document associated with the first hierarchical expertise level. Additionally, by encoding relative levels of domain expertise or other domain-specific relationships in graph edges that indicate cross-domain relations, some embodiments may improve the speed and accuracy of responses to queries for information or provide other aspects of expert guidance. It should be emphasized, though, that not all embodiments necessarily provide these advantages, as there are several independently useful ideas described herein, and some implementations may only apply a subset of these techniques. As used in this disclosure, the term “ontology” may be used interchangeably with the term “ontology graph,” unless otherwise indicated, where an entry of an ontology may include a vertex of the ontology).
As per claim 10, the combination of Bender, Gnanasambandam, and Peng teaches each of the limitations of claims 1 and 5.
In addition, Bender teaches:
determine the connections between the extracted topics and the plurality of nodes by determining, utilizing classifiers trained on the predefined ontology of topics and a plurality of manually labeled text responses, that the topic indicated in the text response corresponds to one or more topics in the predefined ontology of topics (Col. 25 line 62 - col. 26 line 25 teaches transfer parameters of machine learning model, where the parameters may include a set of neural network parameters such as weights, biases, activation function parameters, or other values of the neurons of a neural network. Once transferred, these parameters may be used by a new instance of the neural network model or other machine learning model. For example, some embodiments can train a BERT-based machine learning model to predict answers based on training queries from a stored library of queries and answers, where the answers for the queries may include semantic search results. Some embodiments may train a machine learning model based on a set of training queries and a corresponding set of training documents that should be retrieved when the system is provided with the set of training queries. Col. 50 lines 4 - 20 teach the set of task-specific layers 1214 may include a second set of neural network parameters that are updated with a set of training operations to perform various tasks. In some embodiments, the set of training operations may use parameters obtained from a template task library 1204 to update parameters of the set of task-specific layers 1214. The template task library 1204 may include training task parameters such as a training dataset for tasks such as summarization, text classification, language modeling, named entry recognition, text encoding, ontology lookup, natural language generation, question-answering, text representation, or the like. For example, some embodiments may train the learning model 1210 by importing neural network parameters from the template task library 1204 into a set of neural network layers of the set of task-specific layers 1214 and use a first training dataset of the template task library 1204 to train the modified learning model 1210).
As per claim 12, the combination of Bender, Gnanasambandam, and Peng teaches each of the limitations of claim 11.
In addition, Bender teaches:
utilizing a machine-learning model or keyword matching, (Col. 25 line 62 - Col. 26 line 10 teaches transfer parameters of machine learning model, where the parameters may include a set of neural network parameters such as weights, biases, activation function parameters, or other values of the neurons of a neural network. Once transferred, these parameters may be used by a new instance of the neural network model or other machine learning model. For example, some embodiments can train a BERT-based machine learning model to predict answers based on training queries from a stored library of queries and answers, where the answers for the queries may include semantic search results. Some embodiments may train a machine learning model based on a set of training queries and a corresponding set of training documents that should be retrieved when the system is provided with the set of training queries).
associating the extracted topic with a topic node of the plurality of nodes based on the relationship (Col. 4 lines 31-57 teach such other information can include the document's origin, metadata associated with the document, a data format, or the like. A domain vector for a document can indicate various types of information relevant to the usefulness of the document, such as an associated expertise level for each of a plurality of domains, a count of words, a count of words having more than a specified number of syllables, or the like. Col. 7 line 62 - Col. 8 line 10 teaches store ontology data in a set of SQL tables of the ontology database 138, where each record of the SQL table may represent a vertex record and include, as table fields, parent vertex identifiers, child vertex identifiers, categories indicating relationship types between vertices, scores associated with the relationship category, or the like. Some embodiments may store data in a combination of relational and non-relational databases. Col. 17 lines 55-65 teach categorize a vertex group determined from a clustering method as a first type of vertex group, where vertices of a vertex group of the first type of vertex group may be associated with vectors categorized as being part of a same cluster. In some embodiments, the vertex group may represent a ‘concept’ in a domain, where the concept may be shared amongst multiple classes of the domain).
Bender teaches receiving requests for data and utilizing knowledge graphs to gather, store, and profile data but does not explicitly teach extracting data through survey questions and real-time communications as described by the following citations from Gnanasambandam:
inferring,... a relationship between an extracted topic extracted from the question data and response data for the one or more electronic survey questions and a topic of the predefined ontology of topics based on the relationships between the plurality of nodes (Paragraph Number [0287] teaches receiving a first data including user registration data (block 702); and providing a health assessment and receiving second data including health assessment answers (block 704). In various embodiments, the health assessment is a micro-survey with dynamically formulated questions presented to the user. Paragraph Number [0303] teaches the cognitive agent 110 can answer a library of questions and provide content for many questions a user has as it related to diabetes. The information provided for purposes of educating a user is based on an overall health plan of the user, which is based on meta data analysis of interactions with the user, and an analysis of the education level of the user. Paragraph Number [0304] teaches the user inputs text using the input box 916, which instructs the user to “Talk to me or type your question”.).
A person of ordinary skill would have been motivated to combine these references for the same reasons put forth in regard to claim 1.
As per claim 13, the combination of Bender, Gnanasambandam, and Peng teaches each of the limitations of claims 11 and 12.
In addition, Bender teaches:
determine a first entity ontology comprising a first set of topics associated with a first entity (Col. 55 line 64 - Col. 56 line 14 teaches various types of scoring models may be used to score n-grams of a natural-language text document, where n-grams of a natural-language text document may be scored by individually scoring the n-grams or by scoring text sections including the n-grams. In some embodiments, the scoring model may be a model used for extractive summarization methods, where one or more text sections may be selected as summarizing text sections based on the value of the corresponding scores. Some embodiments may determine topic scores for each sentence of a natural language document. In some embodiments, each respective topic score corresponding to a respective text section and may indicate relevance to a topic, where the topic may be determined from a query or an updated query. For example, a first topic of a query may include the phrase “atrial fibulation” based on the query including the phrase “atrial fibulation,” and a second topic of the query may include the acronym “NVAF” based on a set of cross-graph associations between the n-gram “NVAF” and the n-gram “Atrial Fibulation.”).
determine first correspondences between the first set of topics of the first entity ontology and the predefined ontology of topics (Col. 56 line 50-67 teaches as described elsewhere in this disclosure, some embodiments may generate scores for individual n-grams of a document. For example, some embodiments may generate a score for each n-gram of a sentence of a document, and text sections comprising the sentence may be scored based on the individual scores of the n-grams. In some embodiments, the scoring model may include a neural network that determines a sentence score. For example, some embodiments may use a recurrent neural network to determine a learned representation of a sentence with respect to a specific topic, where different RNNs may be used to determine different sentence scores for a same sentence with respect to different topics. Some embodiments may then select a set of text sections that satisfy a score criteria such as a relevance threshold (e.g., in the form of a minimum score threshold) or select a set of text sections based on their rankings to determine which text sections to analyze or display in a user interface (UI). Col. 72 lines 5-17 teach as disclosed elsewhere in this disclosure, some embodiments may access different indices or different portions and index based on a user context parameter, such as one identifying a domain category value. Alternatively, or in addition, some embodiments may apply different scoring systems based on a user context. For example, some embodiments may use a first scoring model to determine scores for a set of sentences of a document, where the scores may indicate a predicted relevance to a first topic. Some embodiments may then use a second scoring model to determine a different set of scores for the same set of sentences of a document, where the second set of scores may indicate a predicted relevance to a second topic).
determine the relationships between the plurality of nodes based on the first correspondences between the first set of topics of the first entity ontology and the predefined ontology of topics (Col. 56 line 50-67 teaches as described elsewhere in this disclosure, some embodiments may generate scores for individual n-grams of a document. For example, some embodiments may generate a score for each n-gram of a sentence of a document, and text sections comprising the sentence may be scored based on the individual scores of the n-grams. In some embodiments, the scoring model may include a neural network that determines a sentence score. For example, some embodiments may use a recurrent neural network to determine a learned representation of a sentence with respect to a specific topic, where different RNNs may be used to determine different sentence scores for a same sentence with respect to different topics. Some embodiments may then select a set of text sections that satisfy a score criteria such as a relevance threshold (e.g., in the form of a minimum score threshold) or select a set of text sections based on their rankings to determine which text sections to analyze or display in a user interface (UI). Col. 72 lines 5-17 teach as disclosed elsewhere in this disclosure, some embodiments may access different indices or different portions and index based on a user context parameter, such as one identifying a domain category value. Alternatively, or in addition, some embodiments may apply different scoring systems based on a user context. For example, some embodiments may use a first scoring model to determine scores for a set of sentences of a document, where the scores may indicate a predicted relevance to a first topic. Some embodiments may then use a second scoring model to determine a different set of scores for the same set of sentences of a document, where the second set of scores may indicate a predicted relevance to a second topic).
As per claim 14, the combination of Bender, Gnanasambandam, and Peng teaches each of the limitations of claims 11-13.
In addition, Bender teaches:
determine a second entity ontology comprising a second set of topics associated with a second entity (Col. 55 line 64 - Col. 56 line 14 teaches various types of scoring models may be used to score n-grams of a natural-language text document, where n-grams of a natural-language text document may be scored by individually scoring the n-grams or by scoring text sections including the n-grams. In some embodiments, the scoring model may be a model used for extractive summarization methods, where one or more text sections may be selected as summarizing text sections based on the value of the corresponding scores. Some embodiments may determine topic scores for each sentence of a natural language document. In some embodiments, each respective topic score corresponding to a respective text section and may indicate relevance to a topic, where the topic may be determined from a query or an updated query. For example, a first topic of a query may include the phrase “atrial fibulation” based on the query including the phrase “atrial fibulation,” and a second topic of the query may include the acronym “NVAF” based on a set of cross-graph associations between the n-gram “NVAF” and the n-gram “Atrial Fibulation.”).
determine second correspondences between the second set of topics of the second entity ontology and the predefined ontology of topics, the first correspondences associated with the first set of topics being different than the second correspondences associated with the second set of topics based on different terminologies for the first set of topics and the second set of topics (Col. 56 line 50-67 teaches as described elsewhere in this disclosure, some embodiments may generate scores for individual n-grams of a document. For example, some embodiments may generate a score for each n-gram of a sentence of a document, and text sections comprising the sentence may be scored based on the individual scores of the n-grams. In some embodiments, the scoring model may include a neural network that determines a sentence score. For example, some embodiments may use a recurrent neural network to determine a learned representation of a sentence with respect to a specific topic, where different RNNs may be used to determine different sentence scores for a same sentence with respect to different topics. Some embodiments may then select a set of text sections that satisfy a score criteria such as a relevance threshold (e.g., in the form of a minimum score threshold) or select a set of text sections based on their rankings to determine which text sections to analyze or display in a user interface (UI). Col. 72 lines 5-17 teach as disclosed elsewhere in this disclosure, some embodiments may access different indices or different portions and index based on a user context parameter, such as one identifying a domain category value. Alternatively, or in addition, some embodiments may apply different scoring systems based on a user context. For example, some embodiments may use a first scoring model to determine scores for a set of sentences of a document, where the scores may indicate a predicted relevance to a first topic. Some embodiments may then use a second scoring model to determine a different set of scores for the same set of sentences of a document, where the second set of scores may indicate a predicted relevance to a second topic.).
As per claim 15, the combination of Bender, Gnanasambandam, and Peng teaches each of the limitations of claim 11.
In addition, Bender teaches:
associate the topic indicated the response data for the one or more electronic survey questions by connecting the experience node to the topic node via one or more edges indicating a relationship between the response data for the one or more electronic survey questions and the topic node (Col. 18 lines 26-43 teach some embodiments may determine a hierarchy of ontologies based on one or more vertex groups associated with a vector cluster via an edge connection between the vertex group and one or more shared connections. Some embodiments may determine that a first vertex group of a first ontology graph may be associated with a vertex of a second ontology graph via one or more shared vertices. For example, a first vertex group may include a set of vertices corresponding with a first set of learned representations that includes the embedding vector [x.sub.1, x.sub.2, x.sub.3]. Some embodiments may determine that the embedding vector corresponds with a vertex in a second ontology graph, and, in response, determine that the first vertex group is associated with the vertex in the second ontology graph. Additionally, if the vertex in the second ontology graph is part of a second vertex group, some embodiments may determine that the first vertex group is associated with the second vertex group of the ontology graph. (See also Col. 4 lines 31-57 and Col. 7 line 62 - Col. 8 line 10)).
Bender teaches receiving requests for data and utilizing knowledge graphs to gather, store, and profile data but does not explicitly teach extracting data through survey questions and real-time communications as described by the following citations from Gnanasambandam:
generate an experience node comprising experience data of the response data for the one or more electronic survey questions and data inferred from the response data for the one or more electronic survey questions (Paragraph Number [0287] teaches receiving a first data including user registration data (block 702); and providing a health assessment and receiving second data including health assessment answers (block 704). In various embodiments, the health assessment is a micro-survey with dynamically formulated questions presented to the user. Paragraph Number [0303] teaches the cognitive agent 110 can answer a library of questions and provide content for many questions a user has as it related to diabetes. The information provided for purposes of educating a user is based on an overall health plan of the user, which is based on meta data analysis of interactions with the user, and an analysis of the education level of the user. Paragraph Number [0304] teaches the user inputs text using the input box 916, which instructs the user to “Talk to me or type your question”).
A person of ordinary skill would be motivated to combine these references as described in regard to claim 1.
Bender teaches receiving requests for data and utilizing knowledge graphs to gather, store, and profile data but does not explicitly teach generating nodes that include person nodes that represent information about a particular user and containing information about that specific user as described by the following citations from Peng:
generate, in response to associating a topic indicated in the response data for the one or more electronic survey questions with a topic node, a leaf user node connected by an edge to the topic node based on respondent data associated with a user of the one or more client devices, the leaf user node comprising a person node type (Paragraph Number [0076] teaches a user node 602 may correspond to a user of the social-networking system 160 or the assistant system 160. As an example and not by way of limitation, a user may be an individual (human user), an entity (e.g., an enterprise, business, or third-party application), or a group (e.g., of individuals or entities) that interacts or communicates with or over the social-networking system 160 or the assistant system 160. In particular embodiments, when a user registers for an account with the social-networking system 160, the social-networking system 160 may create a user node 602 corresponding to the user, and store the user node 602 in one or more data stores. Paragraph Number [0079] A user viewing the third-party web interface may perform an action by selecting one of the icons (e.g., “check-in”), causing a client system 130 to send to the social-networking system 160 a message indicating the user's action. In response to the message, the social-networking system 160 may create an edge (e.g., a check-in-type edge) between a user node 602 corresponding to the user and a concept node 604 corresponding to the third-party web interface or resource and store edge 606 in one or more data stores. Paragraph Number [0081] teaches although this disclosure describes edges between a user node 602 and a concept node 604 representing a single relationship, this disclosure contemplates edges between a user node 602 and a concept node 604 representing one or more relationships. As an example and not by way of limitation, an edge 606 may represent both that a user likes and has used at a particular concept. Alternatively, another edge 606 may represent each type of relationship (or multiples of a single relationship) between a user node 602 and a concept node 604 (as illustrated in FIG. 6 between user node 602 for user “E” and concept node 604).).
A person of ordinary skill would be motivated to combine these references as described in regard to claim 1.
As per claim 16, the combination of Bender, Gnanasambandam, and Peng teaches each of the limitations of claims 11 and 12.
In addition, Bender teaches:
determining that the first set of data corresponds to a first plurality of nodes connected to the topic node, the first plurality of nodes comprising first information related to a first entity (Col. 18 lines 26-43 teach some embodiments may determine a hierarchy of ontologies based on one or more vertex groups associated with a vector cluster via an edge connection between the vertex group and one or more shared connections. Some embodiments may determine that a first vertex group of a first ontology graph may be associated with a vertex of a second ontology graph via one or more shared vertices. For example, a first vertex group may include a set of vertices corresponding with a first set of learned representations that includes the embedding vector [x.sub.1, x.sub.2, x.sub.3]. Some embodiments may determine that the embedding vector corresponds with a vertex in a second ontology graph, and, in response, determine that the first vertex group is associated with the vertex in the second ontology graph. Additionally, if the vertex in the second ontology graph is part of a second vertex group, some embodiments may determine that the first vertex group is associated with the second vertex group of the ontology graph).
determining that the second set of data corresponds to a second plurality of nodes connected to the topic node, the second plurality of nodes comprising second information related to a second entity (Col. 18 lines 26-43 teach some embodiments may determine a hierarchy of ontologies based on one or more vertex groups associated with a vector cluster via an edge connection between the vertex group and one or more shared connections. Some embodiments may determine that a first vertex group of a first ontology graph may be associated with a vertex of a second ontology graph via one or more shared vertices. For example, a first vertex group may include a set of vertices corresponding with a first set of learned representations that includes the embedding vector [x.sub.1, x.sub.2, x.sub.3]. Some embodiments may determine that the embedding vector corresponds with a vertex in a second ontology graph, and, in response, determine that the first vertex group is associated with the vertex in the second ontology graph. Additionally, if the vertex in the second ontology graph is part of a second vertex group, some embodiments may determine that the first vertex group is associated with the second vertex group of the ontology graph. Col. 23 lines 6-19 teach various other associations in the index may be made, such as associating the first vertex with the second document or the third document. By associating documents of a corpus in an index based on hierarchical associations between vertices of knowledge graphs and vertex adjacency in a knowledge graph, some embodiments may increase the speed of document retrieval by using the index. Additionally, as further described below, some embodiments may generate question-answer pairs based on the knowledge graph and include the question-answer pairs in an index. For example, some embodiments may associate a specific query or type of query with a specific document or set of documents. Some embodiments may include this association representing a question-answer pair in the index).
generating the digital benchmark based on a comparison of first information to the second information (Col. 6 lines 6-33 teach an ontology data model or knowledge graph organized by the ontology data model may improve the relevance of retrieved documents by accounting for a user's domain expertise or specific interests. Such operations may be especially useful in specialized applications where similar concepts may be disclosed in documents at differing levels of domain expertise, differing levels of security classification, or with differing amounts of relevance to subdomains. For example, if a user associated with a first hierarchical expertise level performs a search, some embodiments may obtain a first document associated with the first hierarchical expertise level and a second document associated with a second hierarchical expertise level. Some embodiments may then provide a user with the document associated with the first hierarchical expertise level. Additionally, by encoding relative levels of domain expertise or other domain-specific relationships in graph edges that indicate cross-domain relations, some embodiments may improve the speed and accuracy of responses to queries for information or provide other aspects of expert guidance. It should be emphasized, though, that not all embodiments necessarily provide these advantages, as there are several independently useful ideas described herein, and some implementations may only apply a subset of these techniques. As used in this disclosure, the term “ontology” may be used interchangeably with the term “ontology graph,” unless otherwise indicated, where an entry of an ontology may include a vertex of the ontology).
As per claim 18, the combination of Bender, Gnanasambandam, and Peng teaches each of the limitations of claim 17.
In addition, Bender teaches:
generating a modification to the predefined ontology of topics based on receiving, from an additional client device, additional survey data comprising additional question data and additional response data for one or more additional electronic survey questions (Col. 31 lines 48-67 teaches integrate a first ontology with other ontologies to form the ontology system. Some embodiments may integrate the first ontology with other ontologies based on a second ontology associating words or other n-grams of the first ontology with the other ontologies, allowing words and concepts to be hierarchically linked across different domains or classes of expertise within the domains. Some embodiments may include a set of UI elements to control a modular knowledge system and provide visual indicators to indicate whether an ontology combination passes or fails a set of rules. Some embodiments may present visualizations of the different types of edges governing vertex relationships within an ontology graph or other vertex relationships, where the ontology graph may represent an ontology. Some embodiments may further present visualizations of query interpretations based on the set of ontology combinations. Col. 35 lines 24-57 teach after updating a set of ontology graphs by forming cross-graph edges relating vertices of different ontology graphs, some embodiments may provide previously-undetected associations between different vertices. For example, the first cross-graph edge associating the eighth vertex of the third ontology graph with the ninth vertex of the second ontology graph may be used to determine an association between the n-gram “aFib” and the n-gram “Valvular Atrial Fibrillation.” Additionally, some embodiments may detect previously non-established links between vertices of a same ontology graph by using edges associating vertices of ontology graphs having different domains or classes).
updating the knowledge graph by updating one or more of the person nodes, the activity nodes, the experience nodes, or the operational nodes to according to the modification to the predefined ontology of topics (Col. 9 lines 20-43 teach use of various types of data to generate or otherwise update the ontology data stored in the ontology data repository 230. The data may include a set of existing ontology data 211, a set of natural-language text documents 212, or a set of structured data 214. The existing knowledge graph may be stored in various ways, such as in a relational data structure, and may be imported into the ontology data repository 230. Different data types may be combined to update an ontology data model, such as one stored in an ontology data model record 231. The ontology data model record may store values for record fields such as object categories, relationships between the categories, directional indicators of the relationships, or the like. Alternatively, or in addition, the ontology data model may be stored in a knowledge graph such as a knowledge graph 232, which may be formatted in a specified ontology data model. (Examiner asserts that the relational data structure that combines data types including object categories, relationships, etc., is equivalent to an experience node based upon Paragraph Number [0058] of Applicant's specification which provides: "the survey knowledge graph system 102 determines respondent feedback such as ratings, opinions, sentiment, emotion, or other responses that indicate experiences of users with activities or topics"))
As per claim 19, the combination of Bender, Gnanasambandam, and Peng teaches each of the limitations of claim 17.
In addition, Bender teaches:
generating experience data by comparing the first set of data to the second set of data within the knowledge graph to identify one or more similarities in patterns between the first set of data and the second set of data (Col. 9 lines 20-43 teach use of various types of data to generate or otherwise update the ontology data stored in the ontology data repository 230. The data may include a set of existing ontology data 211, a set of natural-language text documents 212, or a set of structured data 214. The existing knowledge graph may be stored in various ways, such as in a relational data structure, and may be imported into the ontology data repository 230. Different data types may be combined to update an ontology data model, such as one stored in an ontology data model record 231. The ontology data model record may store values for record fields such as object categories, relationships between the categories, directional indicators of the relationships, or the like. Alternatively, or in addition, the ontology data model may be stored in a knowledge graph such as a knowledge graph 232, which may be formatted in a specified ontology data model. (Examiner asserts that the relational data structure that combines data types including object categories, relationships, etc., is equivalent to an experience node based upon Paragraph Number [0058] of Applicant's specification which provides: "the survey knowledge graph system 102 determines respondent feedback such as ratings, opinions, sentiment, emotion, or other responses that indicate experiences of users with activities or topics")).
Bender teaches receiving requests for data and utilizing knowledge graphs to gather, store, and profile data but does not explicitly teach extracting data through survey questions and real-time communications as described by the following citations from Gnanasambandam:
generating, based the experience data, an experience node of plurality of nodes corresponding to the predefined ontology of topics (Paragraph Number [0287] teaches receiving a first data including user registration data (block 702); and providing a health assessment and receiving second data including health assessment answers (block 704). In various embodiments, the health assessment is a micro-survey with dynamically formulated questions presented to the user. Paragraph Number [0303] teaches the cognitive agent 110 can answer a library of questions and provide content for many questions a user has as it related to diabetes. The information provided for purposes of educating a user is based on an overall health plan of the user, which is based on meta data analysis of interactions with the user, and an analysis of the education level of the user. Paragraph Number [0304] teaches the user inputs text using the input box 916, which instructs the user to “Talk to me or type your question”).
A person of ordinary skill would be motivated to combine these references as described in regard to claim 1.
As per claim 20, the combination of Bender, Gnanasambandam, and Peng teaches each of the limitations of claim 17.
In addition, Bender teaches:
generating a modification to the predefined ontology of topics based on receiving, from an additional client device, additional survey data comprising additional question data and additional response data for one or more additional electronic survey questions (Col. 31 lines 48-67 teaches integrate a first ontology with other ontologies to form the ontology system. Some embodiments may integrate the first ontology with other ontologies based on a second ontology associating words or other n-grams of the first ontology with the other ontologies, allowing words and concepts to be hierarchically linked across different domains or classes of expertise within the domains. Some embodiments may include a set of UI elements to control a modular knowledge system and provide visual indicators to indicate whether an ontology combination passes or fails a set of rules. Some embodiments may present visualizations of the different types of edges governing vertex relationships within an ontology graph or other vertex relationships, where the ontology graph may represent an ontology. Some embodiments may further present visualizations of query interpretations based on the set of ontology combinations. Col. 35 lines 24-57 teach after updating a set of ontology graphs by forming cross-graph edges relating vertices of different ontology graphs, some embodiments may provide previously-undetected associations between different vertices. For example, the first cross-graph edge associating the eighth vertex of the third ontology graph with the ninth vertex of the second ontology graph may be used to determine an association between the n-gram “aFib” and the n-gram “Valvular Atrial Fibrillation.” Additionally, some embodiments may detect previously non-established links between vertices of a same ontology graph by using edges associating vertices of ontology graphs having different domains or classes).
updating the knowledge graph by updating the relationships between the plurality of nodes according to the modification to the predefined ontology of topics (Col. 9 lines 20-43 teach use of various types of data to generate or otherwise update the ontology data stored in the ontology data repository 230. The data may include a set of existing ontology data 211, a set of natural-language text documents 212, or a set of structured data 214. The existing knowledge graph may be stored in various ways, such as in a relational data structure, and may be imported into the ontology data repository 230. Different data types may be combined to update an ontology data model, such as one stored in an ontology data model record 231. The ontology data model record may store values for record fields such as object categories, relationships between the categories, directional indicators of the relationships, or the like. Alternatively, or in addition, the ontology data model may be stored in a knowledge graph such as a knowledge graph 232, which may be formatted in a specified ontology data model. (Examiner asserts that the relational data structure that combines data types including object categories, relationships, etc., is equivalent to an experience node based upon Paragraph Number [0058] of Applicant's specification which provides: "the survey knowledge graph system 102 determines respondent feedback such as ratings, opinions, sentiment, emotion, or other responses that indicate experiences of users with activities or topics")).
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to MATTHEW H. DIVELBISS whose telephone number is (571) 270-0166. The fax phone number is 571-483-7110. The examiner can normally be reached on M-Th, 7:00 - 5:00. 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, Jerry O'Connor can be reached on (571) 272-6787.
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.
/MATTHEW H DIVELBISS/Examiner, Art Unit 3624
Prosecution Insights