DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Response to Arguments
Remarks pages 11-16, Applicant contends:
Claims, especially as amended, satisfy requirements for 101.
Response:
The improvement of creating an improved knowledge structure/system for use in queries or inferences is seen as satisfying the requirements for 101. Thus arguments are seen as convincing and 101 rejections are withdrawn.
Remarks page 16-17, Applicant contends:
The amended claim limitations provide clarification on the distinction between domain values as previously interpreted and domain values as recited.
Response:
The amendments to the claim limitations are interpreted as clarifying and further narrowing the interpretation of domain values and thus other aspects of the current invention. The rejections under 103 are updated to fit the updated interpretation as a result of the change in meaning for elements such as domain values.
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.
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, 3-6, 8-11, 13-18 is/are rejected under 35 U.S.C. 103 as being unpatentable over Tung et al (US 20200012738 A1), referred to as Tung in this document, and further in combination of Paulheim ("Knowledge graph refinement: A survey of approaches and evaluation methods"), referred to as Paulheim in this document, and further in combination with Lei et al (US 20220237446), referred to as Lei in this document.
Regarding Claim 1:
Tung teaches:
A method for knowledge extraction from heterogeneous data sources, comprising,
[Tung 0013]: “Second, the knowledge graph is hydrated with information by ingesting knowledge from multiple data sources and different knowledge extraction techniques [A method for knowledge extraction from heterogeneous data sources comprising] (e.g., natural language processing (NLP), schema mapping, computer visions, or the like) to create the vertices and edges in the knowledge graph. Each data source may create its own data processing pipeline for extracting data to include into the knowledge graph being constructed.”
extracting, by a computing device, raw data from a first data source
[Tung 0020]: “Referring back to the processing engine 115 shown in FIG. 1, the processing engine 115 is configured to extract data from different data sources in response to a received information query. For example, the processing engine 115 may apply F techniques on data to obtain an intent, sentiment, and/or nuances and context of spoken sentences of data to more accurately extract pertinent data. The processing engine 115 may further apply structured data denormalization and normalization techniques when parsing the data from the different data sources to better extract data [extracting, by a computing device, raw data from a first data source]. The processing engine 115 may further apply computer vision/object detection and recognition when parsing the data from the different data sources to better extract data. The techniques applied by the processing engine 115 to extract data from the different data sources may be considered to be the first step in the knowledge graph hydration process.”
Tung teaches the above limitation, but an alternative teaching for reference in regards to the currently updated interpretation of knowledge models and domain values is provided in the teachings of Lei.
synthesizing, by the computing device, a first set of entities based on the extracted raw data
[Tung 0018]: “The processing layer 110 also includes a pipeline repository 114 that stores data extraction rules, techniques, and protocols for extracting data from different data sources. The pipeline repository 114 stores a repository of instruction code and instruction protocols that can be used to extract entities [synthesizing, by the computing device, a first set of entities based on the extracted raw data as getting entities from data using instructions is interpreted as synthesizing entities from the data under BRI] and other information from a corpus of data. A knowledge graph storage 131 stores initial graph ontologies (i.e., graph schemas), where a graph ontology may be domain specific or otherwise customized for particular applications. A graph ontology is a high-level schema defining how various entities are related. For example, the graph ontology includes the information for defining all entity types, edge types, and their hierarchies for a specific knowledge graph. A graph ontology may be referred to as a graph schema within this disclosure.”
Tung teaches the above limitation, but an alternative teaching for reference in regards to the currently updated interpretation of knowledge models and domain values is provided in the teachings of Lei.
extracting, by the computing device, a second set of entities from a second data source
mapping the extracted second set of entities to the extracted one or more concepts, the extracted one or more connectors, and the identified relations, based on the generated first knowledge structure, wherein the second data source is associated with the first data source
[Tung 0024]: “A new data processing pipeline may be onboarded to the KDMS 1 by, for example, accessing a new data source. A new data processing pipeline may also be onboarded to the KDMS 1 by, for example, identifying new entity relationships. FIG. 3 shows two exemplary data processing pipelines 300, as well as a graphical representation of the orchestration process. A first data processing pipeline A receives a newsfeed as an input, and outputs a Politician entity type as an output (which may, or may not, be empty). A second data processing pipeline B includes the Politician entity type as an input, and the President entity type as an output [mapping the extracted second set of entities to the extracted one or more concepts, the extracted one or more connectors, and the identified relations,]. Both the first data processing pipeline A and the second data processing pipeline B may be running NLP to extract information [extracting, by the computing device, a second set of entities from a second data source] from their respective data sources.”
Support for the connection to the first knowledge structure being a form of graph ontology/schema for the new interpretation resulted from the meaning of domain value being narrowed is supported in Tung 13 along with other information detailing how pipelines and such are related to the extracting of information from a second source to map information to a knowledge graph.
[Tung 0013]: “Constructing a knowledge graph may include two steps. First, a graph schema [based on the generated first knowledge structure,] definition is obtained for the knowledge graph and refinement is applied as the knowledge graph is being generated. This defines the types of vertices and edges that are generated into the knowledge graph [wherein the second data source is associated with the first data source as for the ontology/schema being made from the first data source (as noted by the claims and shown in associated mapping) being usable for the knowledge graph that is filled in with data from another source(s) shows an association as the relationships or concepts from the first source are being identified within the second, thus showing association or overlap]. Second, the knowledge graph is hydrated with information by ingesting knowledge from multiple data sources and different knowledge extraction techniques (e.g., natural language processing (NLP), schema mapping, computer visions, or the like) to create the vertices and edges in the knowledge graph. Each data source may create its own data processing pipeline for extracting data to include into the knowledge graph being constructed.”
converting, by the computing device, the mapping of the extracted second set of entities into one or more data structures
generating, by the computing device, a second knowledge structure from the converted one or more data structures
[Tung 0022]: “Each intermediary result may be data that comprises a portion [converting, by the computing device, the mapping of the extracted second set of entities into one or more data structures] of the knowledge graph being constructed [generating, by the computing device, a second knowledge structure from the converted one or more data structures]. An intermediary result data stored on the staging repository may be further refined or ingested into the knowledge graph being constructed when the processing engine 115 determines further refinement is not needed.”
updating, by the computing device, a knowledge graph store based on the generated second knowledge structure
[Tung 0023]: “The integration layer 120 implements the orchestration process via orchestration circuitry 123 for determining whether intermediary result data stored on the staging repository can be further refined. The further refinement of intermediary results may include the addition of a next level (i.e., deeper) sub-concept to the knowledge graph being constructed by an additional processing step. The KDMS 1 maintains a record [updating, by the computing device, a knowledge graph store based on the generated second knowledge structure] of expected input and output types, in terms of concepts and relationships, for each data processing pipeline. These records may be stored as pipeline metadata 122. Records are created when new data processing pipelines are onboarded to the KDMS 1.” This notes that data from the processes is added to the records.
Tung Figure 7 also notes storing knowledge graph data on part 708 for an alternative mapping of storing knowledge graphs.
Further support of storage for knowledge graphs is in paragraph 37 of Tung noting “The KDMS 1 further includes the graph layer 130 comprised of the knowledge graph storage 131. The knowledge graph storage 131 stores entities (nodes), relationships (edges), and attributes (node/edge properties).”
And querying, by the computing device, the updated knowledge graph store based on a received query to retrieve entity relationships responsive to the received query
[Tung 0012]: “To address these technical problems, a knowledge graph is disclosed that offers an innovative data structure that presents relevant information in response to a data query [And querying, by the computing device, the updated knowledge graph store based on a received query to retrieve entity relationships responsive to the received query where details about a knowledge graph store are taught earlier in claim 1], as well as relationship information between the relevant information. The knowledge graph includes a knowledge base of relevant information that is structured in a graph presentation that captures entities (i.e., nodes), relationships (i.e., edges), and attributes (i.e., node properties or edge properties) with semantic meaning. This graph data structure model offered by the knowledge graph provides the semantic meaning of the included data, by modeling data with an ontology or taxonomy. Accordingly, technical improvements are realized when a computing device structures information into knowledge graphs and runs search queries on the knowledge graphs, which specifically result in the retrieval of more relevant and accurate information, in a shorter amount of time.”
Tung does not explicitly teach:
extracting, by the computing device, one or more concepts and one or more connectors by classifying each entity of the synthesized first set of entities as a concept or a connector, based on a pre-existing knowledge model, wherein the pre-existing knowledge model comprises a set of concepts, a set of connectors, and domain values associated with the set of concepts and the set of connectors, wherein each domain value of the domain values is an instance value associated with a respective concept or connector
Tung teaches aspects related to extracting concepts and connectors, but not in the same way described by the current claims with current limitations describing domain values. Previous office action had a broader meaning of domain values and thus knowledge models to work with. As noted by the specification, a knowledge model is some form of knowledge structure or database that contains knowledge such as concepts and connections [Current Invention 0045]. The change to domain values to require “each domain values…is an instance value associated with a respective concept or connector” narrows the relationship of the domain values with parts of the knowledge model. Previous limitations only required some form of association, but the current limitations require instances of concepts or connects being domain values, which means multiple domain values can exist per concept or connector and must be an instance of or example of rather than a mere association. The amendment to domain value also altered the interpretation of the pre-existing knowledge model and first knowledge model, as the models are connected to the domain values. The change to domain values results in the knowledge models taking on an interpretation much more like a graph ontology or graph schema. As a result, prior art teaching such elements and how Tung fits in is explained in the current office action.
The extracting, classifying, and such is noted by [Tung 0013]: “Second, the knowledge graph is hydrated with information by ingesting knowledge from multiple data sources and different knowledge extraction techniques (e.g., natural language processing (NLP), schema mapping, computer visions, or the like) to create the vertices and edges in the knowledge graph. Each data source may create its own data processing pipeline for extracting data to include into the knowledge graph being constructed.”. Aspects related to graph ontologies or schemas is mentioned in places such as [Tung 0018], but creating graph ontologies as described by the claims as currently amended does not appear to be in Tung’s teachings.
identifying, by the computing device, relations between the extracted one or more concepts and the one or more connectors, based on the first data source and the pre-existing knowledge model
Tung 0024 notes aspects related to identifying relations between entities or concepts [Tung 0024]: “A new data processing pipeline may be onboarded to the KDMS 1 by, for example, accessing a new data source. A new data processing pipeline may also be onboarded to the KDMS 1 by, for example, identifying new entity relationships...”
generating, by the computing device, a first knowledge structure based on the identified relations
the mapping includes calculating respective distances between each of the second set of entities and one of the extracted one or more concepts or one of the extracted one or more connectors
Paulheim teaches:
the mapping includes calculating respective distances between each of the second set of entities and one of the extracted one or more concepts or one of the extracted one or more connectors
[Paulheim 5.2.1 page 9 of pdf]: "Apriosio et al. [4] use types of entities in different DBpedia language editions (each of which can be understood as a knowledge graph connected to the others) as features for predicting missing types. The authors use a k-NN classifier with different distance measures [the mapping includes calculating respective distances between each of the second set of entities and one of the extracted one or more concepts or one of the extracted one or more connectors] (i.e., kernel functions), such as the overlap of two articles’ categories. In their setting, a combination of different distance measures is reported to provide the best results."
One of ordinary skill in the art, prior to the effective filing date, would have been motivated to combine Tung and Paulheim. Tung and Paulheim are in the same field of endeavor of knowledge graphs or knowledge models. One of ordinary skill in the art would have been motivated to combine Tung and Paulheim to utilize a distance measure or calculation in order to increase coverage or information in a knowledge graph ([Paulheim 5 page 8 of pdf]: "Completion of knowledge graphs aims at increasing the coverage of a knowledge graph. Depending on the target information, methods for knowledge graph completion either predict missing entities, missing types for entities, and/or missing relations that hold between entities").
For clarity of record and to show the combination with Lei would be sensible with the teachings from Paulheim, Lei also indicates utilizing aspects related to distance for mapping in paragraphs 27, 70, and 78. As a result, the teachings of Paulheim would not be unconventional with Lei.
Lei teaches:
extracting, by a computing device, raw data from a first data source
synthesizing, by the computing device, a first set of entities based on the extracted raw data
extracting, by the computing device, one or more concepts and one or more connectors by classifying each entity of the synthesized first set of entities as a concept or a connector, based on a pre-existing knowledge model,
[Lei 0094]: “In some embodiments, the ontology framework engine 804 maps constructs of relational database 814 [extracting, by a computing device, raw data from a first data source] to nodes of a new ontology using the names of constructs of the relational database (e.g., names of tables, columns, etc.) as the names of constructs of the ontology. In some embodiments, the concept augmentation module 806 adds an instance-level concept [synthesizing, by the computing device, a first set of entities based on the extracted raw data] to the new ontology for each distinct value in the relational tables of the relational database 814. The ontology generator 802 connect these new concepts to the nodes created by the ontology framework engine 804 via new “instance of” relationships [extracting, by the computing device, one or more concepts and one or more connectors by classifying each entity of the synthesized first set of entities as a concept or a connector]. This enriches the new ontology with available information from the relational database relational database 814 and allows the ontology generator 802 to bootstrap seed concept matching between two ontologies [based on a pre-existing knowledge model] using exact string matching algorithms. Other approximate string matching algorithms (e.g., edit-distance-based or embeddings-based algorithms) can be used as well, depending on the accuracy requirement.”
Other aspects in Lei support the premise of extracting elements from data sources such as paragraphs 69 and 70, which go over aspects about analyzing and pre-processing text and such from documents.
wherein the pre-existing knowledge model comprises a set of concepts, a set of connectors, and domain values associated with the set of concepts and the set of connectors, wherein each domain value of the domain values is an instance value associated with a respective concept or connector
identifying, by the computing device, relations between the extracted one or more concepts and the one or more connectors, based on the first data source and the pre-existing knowledge model
generating, by the computing device, a first knowledge structure based on the identified relations
[Lei 0095]: “In some embodiments, the neighbor augmentation module 808 adds edges among the nodes of the new ontology created by the ontology framework engine 804 and the concept augmentation module 806. Specifically, if two concepts i and j of an external ontology [wherein the pre-existing knowledge model comprises a set of concepts, a set of connectors, and domain values associated with the set of concepts and the set of connectors, wherein each domain value of the domain values is an instance value associated with a respective concept or connector where domain values being instances of a connector or concept is given by Lei 94 noting “instances of” and [Lei 0024] indicating that ontologies can be enriched by instance level information] (e.g., target ontology from second knowledge base 524 of FIG. 5) have an edge while their counterparts i.sub.0 and j.sub.0 in the new ontology do not [additional support for extracting… based on a pre-existing knowledge model as this quote shows a pre-existing model is used to extract relationships and such for the being created knowledge model], then the neighbor augmentation module 808 adds an edge between i.sub.0 and j.sub.0 in the new ontology. Thus, in some embodiments, the neighbor augmentation module 808 searches the external ontology to identify pairs of concepts that are connected and that correspond with a pair of concepts in the new ontology that lack a connection [identifying, by the computing device, relations between the extracted one or more concepts and the one or more connectors, based on the first data source and the pre-existing knowledge model where the first data source was used to create the new ontology, thus the new ontology has relations both from the first data source and the pre-existing knowledge model (aka external ontology)]. When the neighbor augmentation module 808 identifies such concepts in the new ontology, the neighbor augmentation module 808 augments the new ontology to include a connection between the pair of concepts. The goal is to fill the semantic gap between the new and external ontologies by adding missing structural information. The ontology generator 802 then stores the new ontology [generating, by the computing device, a first knowledge structure based on the identified relations] in electronic memory, such as ontology storage 810 and provides the new ontology to an OntoGNN 812 for processing, for example part of a response to the update request from the requesting system.”
Paragraph 5 of Lei notes a description of a graph ontology having things such as concepts and relationships, thus showing the creation of a graph ontology is akin to creating a knowledge model.
One of ordinary skill in the art, prior to the effective filing date, would have been motivated to combine Tung and Lei. Tung and Lei are in the same field of endeavor of knowledge models and knowledge graphs. One of ordinary skill in the art would have been motivated to combine Tung and Lei in order to help standardize terminology and discover relationships ([Lei 0092]: “In some embodiments, it is desirable to match elements of a relational database 814 (such as tables and columns) to an external ontology (e.g., target ontology from second knowledge base 524 of FIG. 5) to standardize terminology and vocabularies, to discover additional relationships, and support new queries to enable semantically rich applications. In some embodiments, an ontology is created from a relational database by using only the metadata information. Although these ontologies capture schema-level details of the underlying data, they are not as semantically rich as the standard ontologies created by experts.”). Tung notes using ontologies or schemas to help create knowledge graphs in paragraphs 13 and 18, thus the combination with Tung is conventional.
Regarding Claim 3:
The method of claim 1 is taught by Tung, Paulheim, and Lei.
Lei teaches:
wherein the pre-existing knowledge model is are augmented with the generated first knowledge structure
[Lei 0027]: “The overall architecture of the OntoGNN inputs source and target ontologies from a first knowledge base and a second knowledge base.. The predictive matching module minimizes the contrastive matching loss to let the representations of matching concepts have a very small distance while those of unmatched concepts have a large distance. In the illustrated embodiment, the predictive matching module outputs the distances for concept pairs as an output vector. In some embodiments, the output vector is used to create or update a mapping file [wherein the pre-existing knowledge model is are augmented with the generated first knowledge structure where emphasis was added in the quote to show that an update for an ontology was being created where the beginning of the quote notes that the input for this update was two ontologies (a source and target ontologies)] that contains classes or concepts from external database systems or ontologies indexed to identical, similar, or related terms in an internal or known system.”
Lei 0073 also indicates that ontologies are updated, such by getting information that used to make an ontology (which would fit with the mappings in claim 1 as that would mean the ontology created in claim 1 is being used here as part of the system updating ontologies) or by receiving an ontology ([Lei 0073]: “The ontologies provided by the ontology web service 414 are updated [wherein the pre-existing knowledge model is are augmented with the generated first knowledge structure] from time to time, or on demand in response to a request from the application 402, by the ontology updating system 416. For example, in some embodiments, the application 402 provides a knowledge base, such as an ontology or a relational database, to the ontology web service 414 that has concepts that overlap with an ontology maintained by the ontology web service 414. The ontology web service 414 then issues a request to the ontology updating system 416, which responds as described in connection with embodiments disclosed herein. For example, in the case where the ontology web service 414 receives an ontology from the application 402, the ontology updating system 416 generates a map between the ontology received from the application 402 and the ontology maintained by the ontology web service 414. In the case where the ontology web service 414 receives a knowledge base in a format other than an ontology, such as a relational database, the ontology updating system 416 generates a new ontology from the knowledge base, and then generates a map between the new ontology and the ontology maintained by the ontology web service 414.”)
Another possible alternative for showing the updating of a pre-existing knowledge model can be from paragraph 93 and 95 of Lei also indicate that the new ontology made is stored for future use, which as the new ontology is made using elements of a previous ontology (shown in Lei 0095), this stored ontology could be considered an augmented variation of the pre-existing ontology and being stored means that the ontology can take the place as a pre-existing ontology in the future.
One of ordinary skill in the art, prior to the effective filing date, would have been motivated to combine Tung and Lei. Tung and Lei are in the same field of endeavor of knowledge models and knowledge graphs. One of ordinary skill in the art would have been motivated to combine Tung and Lei in order to update ontologies to ensure something related to a metric related to the health and status of a service is not in a non-optimal condition ([Lei 0065]: “In some embodiments, service registry 308 maintains information about the status or health of each service instance including performance information associated each of the service instances. In some such embodiments, such information may include various types of performance characteristics of a given service instance (e.g., cache metrics, etc.) and records of updates. For example, from time to time an administrator updates the ontology web service 310 using an ontology updating system 312. As described in more detail herein, ontology updating system 312 updates ontologies available through the ontology updating system 312 in various ways, including performing an ontology mapping process that discovers connections between concepts of different ontologies that are within a common domain.”). A motivation to combine Tung with Lei is also given in claim 1.
Regarding Claim 4:
The method of claim 1 is taught by Tung, Paulheim, and Lei.
Tung teaches:
further comprising identifying, by the computing device, a set of values related to the extracted one or more concepts and the one or more connectors from the first data source
[Tung 0021]: “For an example of name entity extraction from a newspaper, in such context, politician entities that are extracted from the newspaper may be considered the intermediary result. Now some politician entities that are extracted may be further processed and classified as a specific politician, e.g., President. Then the President and the remaining set of Politicians from the intermediary results are considered to be the final results.”
[Tung 0025]: “The integration layer 120 further implements resolution processing of the intermediary results via resolution circuitry 124. Resolution processing observes the intermediary results and attempts to resolve different expressions of a same entity with information obtained from external data sources [further comprising identifying, by the computing device, a set of values related to the extracted one or more concepts and the one or more connectors from the first data source] 126. For example, FIG. 4 illustrates an exemplary Entity 400 where three different expressions may be attempting to define the same Entity (44th President of the United States, President Obama, Obama). To resolve such situations where different expressions are found for a same real-world entity, the resolution circuitry 124 executes strategies with available contextual information to resolve the different expressions attributed to the same real-world entities.”
[Current Invention 0014]: “Each possible value in a domain is an instance of the concept or the connector. For instance large, medium, small maybe the domain for the concept 'size'; Adam, Smith, Sriram, or employee numbers maybe the domain for the concept 'customer'.”
The example idea of what a value is from Current Invention 0014 is present in Tung 0025. Tung 0025 shows that some of what Tung calls entities are considered values by Current Invention (ex. Obama is an instance of the concept 44th President of the United States).
Lei teaches:
Notes of teachings from Lei are included for possible other interpretation and clarity of record.
further comprising identifying, by the computing device, a set of values related to the extracted one or more concepts and the one or more connectors from the first data source Here this note of teaching from Lei is included in case the domain values being instances of concepts or connectors is intended to be related to the finding of values for concepts and connectors
[Lei 0094]: In some embodiments, the concept augmentation module 806 adds an instance-level concept to the new ontology for each distinct value in the relational tables of the relational database 814. The ontology generator 802 connect these new concepts to the nodes created by the ontology framework engine 804 via new “instance of” relationships [further comprising identifying, by the computing device, a set of values related to the extracted one or more concepts and the one or more connectors from the first data source]. This enriches the new ontology with available information from the relational database relational database 814 and allows the ontology generator 802 to bootstrap seed concept matching between two ontologies using exact string matching algorithms.”
The motivation to combine with Lei is given in claim 1.
Regarding Claim 5:
The method of claim 1 is taught by Tung, Paulheim, and Lei.
Tung teaches:
further comprising removing, by the computing device, noise from the first data source for extracting the one or more concepts and the one or more connectors
[Tung 0026]: “When the resolution processing fails to resolve the entities properly due to lack of available contextual information, the resolution circuitry 124 may perform analytical queries on the knowledge graph to generate a candidate pair of entities to be pruned [further comprising removing, by the computing device, noise from the first data source for extracting the one or more concepts and the one or more connectors] or merged, and calculate a similarity using common connected entities of the candidate entity pair.”
Noise is noted by the current invention to be data that can be construed as ambiguous ([Current Invention 0040]: “In an embodiment, the unrequired data which could be a concept, connector or part of a concept or connector's domain which may have been extracted by the data extractor (201) or classified as an entity by the synthesizer (202) is flagged as a blacklisted word or noise (204). Noise may also include any ambiguous data, or any other unrequired text which is not relevant for the knowledge model or the knowledge graph. This data may be used to retrain the data extractor (201) component in the form of refinements through natural language or regular expressions and other related technologies.”), which is akin to data that difficult to resolve or lacks context as stated in [Tung 0026].
For clarity of record, Lei also notes noise removal in paragraph 70. Motivation to combine with Lei is given in claim 1.
Regarding Claim 6:
Tung teaches:
comprising a processor and a memory comprising instructions executable by the processor to cause the system to perform operations
[Tung 0041]: “As just one example, the system circuitry 604 may include one or more instruction processor 618 and memory 620 [comprising a processor and a memory comprising instructions executable by the processor to cause the system to perform operations].”
The rest of this claim is analogous to claim 1.
Regarding Claim 8:
The cognitive platform of claim 6 is taught by Tung, Paulheim, and Lei.
This claim is analogous to claim 3.
Regarding Claim 9:
The cognitive platform of claim 6 is taught by Tung, Paulheim, and Lei.
This claim is analogous to claim 4.
Regarding Claim 10:
The cognitive platform of claim 6 is taught by Tung, Paulheim, and Lei.
This claim is analogous to claim 5.
Regarding Claim 11:
Tung teaches:
A non-transitory computer readable medium with instructions stored thereon that, when executed by a processor, cause the processor to perform operations
[Tung Claim 18]: “A system comprising: a machine-readable medium, other than a transitory signal [A non-transitory computer readable medium]; and instructions stored on the machine-readable medium that, when executed by processing circuitry [with instructions stored thereon that, when executed by a processor, cause the processor to perform operations]”
The rest of this claim is analogous to claim 1.
Regarding Claim 13:
The cognitive platform of claim 12 is taught by Tung, Paulheim, and Lei.
This claim is analogous to claim 3.
Regarding Claim 14:
The cognitive platform of claim 11 is taught by Tung, Paulheim, and Lei.
This claim is analogous to claim 4.
Regarding Claim 15:
The cognitive platform of claim 11 is taught by Tung, Paulheim, and Lei.
This claim is analogous to claim 5.
Regarding Claim 16:
The method of claim 1 is taught by Tung, Paulheim, and Lei.
Tung teaches:
the one or more data structures include one or more tuples
[Tung 0034]: “Where X and Y are concepts and ->α is the relationship, all defined in a schema. Tp represents the pattern tuple [the one or more data structures include one or more tuples] for X ->α Y. The relationships α’s strength is evaluated by the following two formulations…”
and updating the knowledge graph store includes at least one of inserting the one or more tuples as one or more nodes or one or more edges in a graph store or updating one or more nodes or one or more edges in the graph store to correspond to the one or more tuples
Updating is taught by the updating limitation in claim 1 by Tung 23 [and updating the knowledge graph store includes] and 37. Tung 0034 shows that tuples [at least one of inserting the one or more tuples] are used and thus the updating using tuples is taught by the knowledge a tuple can represent the elements being stored in the knowledge graph store, especially when known in the context of Tung 0037 noting that a knowledge graph is composed of the wanted elements of edges and nodes [Tung 0037]: “The KDMS 1 further includes the graph layer 130 comprised of the knowledge graph storage 131. The knowledge graph storage 131 stores entities (nodes), relationships (edges), [as one or more nodes or one or more edges in a graph store or updating one or more nodes or one or more edges in the graph store to correspond to the one or more tuples] and attributes (node/edge properties).”
Regarding Claim 17:
The system of claim 6 is taught by Tung, Paulheim, and Lei.
This claim is analogous to claim 16.
Regarding Claim 18:
The computer readable medium of claim 11 is taught by Tung, Paulheim, and Lei.
This claim is analogous to claim 16.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
“Knowledge graph embedding with concepts” by Guan et al presents a disclosure that notes connecting entities to concepts while noting that doing so can improve understanding of concepts that use the same word, such “apple” being both a fruit and a company.
Method and Apparatus for Generating Knowledge Graph by Kim et al (US 20220156468 A1) notes the idea of removing noise in relation to concepts and relationships.
US 20190087755 A1 by Hull et al notes having a human in the loop to assist the knowledge graph creation. The specification of the current invention also notes this idea.
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to CHRISTOPHER D DEVORE whose telephone number is (703)756-1234. The examiner can normally be reached Monday-Friday 7:30 am - 5 pm EST.
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, Michael J Huntley can be reached at (303) 297-4307. 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.
/C.D.D./Examiner, Art Unit 2129
/MICHAEL J HUNTLEY/Supervisory Patent Examiner, Art Unit 2129
Prosecution Insights