Abstract: In an example embodiment, a system is provided that takes a knowledge graph as input and generates multiple question templates based on predefined query types. These query types may be, for example, SPARQL query types and may also be configurable. The system repeatedly presents a single randomly or pseudo-randomly picked question template to a user, and asks the user to type a corresponding natural language question. The system persists the answer provided by the user with the question in a specialized format. Then, for each typed question, the system creates additional linguistic variations and asks users to rate them.

Abstract: Techniques and solutions are provided for improved data modelling, including techniques that facilitate data model comparison, integration, or deployment. A core data model is created that has a plurality of elements, which can represent semantic concepts, including relationships between semantic concepts. A plurality of domain models are mapped to core data model elements. In particular, when a domain model element is to be created, the element is mapped to a core data model element. When a relation is to be established between two domain model elements, the relation is mapped to a relation type represented in the core data model. Mapping to a core data model can be enforced as part of domain model creation, allowing different domain models to be compared, since domain elements in different domain models that represent a common semantic will be mapped to the same core data model element.

Abstract: A computer-implemented method includes representing a plurality of database tables as respective vectors in a multi-dimensional vector space, receiving an indication that a first database table represented by a first vector and a second database table represented by a second vector are related to each other, moving positions of the respective vectors representing the plurality of database tables in the multi-dimensional vector space in response to the indication, and grouping the plurality of database tables into one or more table clusters based on positions of the respective vectors representing the plurality of database tables in the multi-dimensional vector space.

Abstract: Properties for an ontology, such as used for semantic query execution, automated analytical reasoning, or for machine learning, are determined using instance graphs. A corpus of documents is received, representing a plurality of domain instances of a domain. Instance graphs are generated for instances of the plurality of instance graphs to provide a plurality of instance graphs. Properties represented in the plurality of instance graphs are determined. At least a portion of the properties are assigned to an ontology for the domain.

Abstract: Class definitions for an ontology of a domain are determined using a materialized instance graph, where the ontology is used for semantic query execution, automated analytical reasoning, or for machine learning. A plurality of instances graphs for a respective plurality of domain instances are received. A materialized instance graph is generated from the plurality of instance graphs. One or more communities represented in the materialized instance graph are determined. Properties associated with respective communities of the one or more communities are determined. Class definitions are generated, where a class corresponds to a community of the one or more communities and at least a portion of properties associated with the community. Class definitions are assigned to the ontology for the domain.

Abstract: Aspects relate to a computer implemented method, computer-readable media and a computer system for detecting data leakage and/or detecting dangerous information. The method comprises receiving a knowledge graph and extracting data from at least one network service. The method further comprises identifying statements in the extracted data. For each identified statement, the method further comprises determining whether the identified statement is public or private using the knowledge graph, and/or determining whether the identified statement is true or false using the knowledge graph.

Abstract: The present disclosure provides techniques and solutions for automatically and dynamically supplementing user prompts to large language models with information to be used by the large language model in formulating a response. In particular, entities are identified in the original prompt. A semantic framework is searched for information about such entities, and such information is added to the original user prompt to provide a modified user prompt. In a particular example, the identified entities comprise triples, and verbalized triples are added to provide the modified user prompt. The modified prompt may be hidden from the user, so that a response of the large language model appears to be in response to the original prompt.

Abstract: A computer-implemented method includes representing a plurality of database tables as respective vectors in a multi-dimensional vector space, receiving an indication that a first database table represented by a first vector and a second database table represented by a second vector are related to each other, moving positions of the respective vectors representing the plurality of database tables in the multi-dimensional vector space in response to the indication, and grouping the plurality of database tables into one or more table clusters based on positions of the respective vectors representing the plurality of database tables in the multi-dimensional vector space.

Abstract: Techniques and solutions are provided for improved data modelling, including techniques that facilitate data model comparison, integration, or deployment. A core data model is created that has a plurality of elements, which can represent semantic concepts, including relationships between semantic concepts. A plurality of domain models are mapped to core data model elements. In particular, when a domain model element is to be created, the element is mapped to a core data model element. When a relation is to be established between two domain model elements, the relation is mapped to a relation type represented in the core data model. Mapping to a core data model can be enforced as part of domain model creation, allowing different domain models to be compared, since domain elements in different domain models that represent a common semantic will be mapped to the same core data model element.

Abstract: Intelligent annotation of data models can be implemented. In one embodiment, the method can receive a data model including entities and relationships between the entities. An entity can include a set of attributes. The method can annotate the data model by defining a logical entity including one or more of the entities, validate the logical entity, and automatically generate a message structure corresponding to the logical entity. The message structure can include properties mapped to at least some of the attributes of entities contained in the logical entity. The properties can be defined in one or more database tables represented by the logical entity.

Abstract: A computer-implemented method can specify a source version and a target version of an application programming interface (API) and a target programming language; and retrieve a difference graph connecting from a source knowledge graph characterizing the source version of the API to a target knowledge graph characterizing the target version of the API. The difference graph includes one or more revision edges representing changes of the API between the source version and the target version. The method can install one or more function packages written in the target programming language and associated with the one or more revision edges; and run the one or more function packages to update the API from the source version to the target version.

Abstract: A method, a system and a computer program for providing data from a directed graph to a language model are provided. The method comprises defining a plurality of conditions and a plurality of patterns, wherein each of the conditions has at least one corresponding pattern. The method further comprises receiving a subset of the directed graph, wherein the subset of the directed graph includes a plurality of statements, wherein each of the statements includes a subject, an object and a predicate relating the subject to the object. The method further comprises for each of the statements in the subset of the directed graph, performing the following: when one of the conditions matches a respective statement and the pattern corresponding to the condition can be applied to the respective statement, computing a string from the respective statement using the pattern.

Abstract: A computer-implemented method can include generating a first knowledge graph from a first version of an application programming interface (API), generating a second knowledge graph from a second version of the API, identifying changes from the second knowledge graph to the first knowledge graph, and generating a difference graph based on the identified changes from the second knowledge graph to the first knowledge graph. The difference graph connects the second knowledge graph to the first knowledge graph via one or more revision edges, which represent the identified changes from the second knowledge graph to the first knowledge graph.

Abstract: A computer-implemented method can include registering a new version of an application programming interface (API) with an API registry, transforming the new version of the API to a first knowledge graph, comparing the first knowledge graph to a second knowledge graph to determine a difference graph, and sending the difference graph to selected entities who have subscribed to the API registry. The second knowledge graph can be transformed from a prior version of the application programming interface. The difference graph connects the second knowledge graph to the first knowledge graph and identifies changes from the second knowledge graph to the first knowledge graph.

Abstract: In an example embodiment, a system is provided that takes a knowledge graph as input and generates multiple question templates based on predefined query types. These query types may be, for example, SPARQL query types and may also be configurable. The system repeatedly presents a single randomly or pseudo-randomly picked question template to a user, and asks the user to type a corresponding natural language question. The system persists the answer provided by the user with the question in a specialized format. Then, for each typed question, the system creates additional linguistic variations and asks users to rate them.

Abstract: Techniques and solutions are provided for improved data modelling, including techniques that facilitate data model comparison, integration, or deployment. A core data model is created that has a plurality of elements, which can represent semantic concepts, including relationships between semantic concepts. A plurality of domain models are mapped to core data model elements. In particular, when a domain model element is to be created, the element is mapped to a core data model element. When a relation is to be established between two domain model elements, the relation is mapped to a relation type represented in the core data model. Mapping to a core data model can be enforced as part of domain model creation, allowing different domain models to be compared, since domain elements in different domain models that represent a common semantic will be mapped to the same core data model element.

Abstract: Various examples are directed to systems and methods for utilizing relationship data in a computing system. The computing system may extract first relationship data from a document and determine a first confidence value describing the first relationship data. The computing system may write the first relationship data to a knowledge graph data structure. The computing system may serve a first user interface page to a user computing device associated with a first user and receive feedback data describing an accuracy of the first relationship data. The computing system may modify a first confidence subunit of a triple data unit associated with the relationship to describe an updated confidence value based on the feedback data and a trust score of the first user.

Abstract: Various examples are directed to systems and methods for utilizing relationship data. A computing system may receive a time-dependent query against a knowledge graph data structure. The computing system may access confirmation data from the knowledge graph data structure, the confirmation data describing a first plurality of confirmation points-in-time at which the first test relationship is true. The computing system may determine that at least one of a beginning or an end of a first time period associated with a test relationship of the time-dependent query is not defined by the knowledge graph data structure. The computing system may determine a response to the first time-dependent query indicating a veracity of the test relationship at a test point-in-time using the first plurality of confirmation points-and-time.

Abstract: A computer-implemented method includes representing a plurality of database tables as respective vectors in a multi-dimensional vector space, receiving an indication that a first database table represented by a first vector and a second database table represented by a second vector are related to each other, moving positions of the respective vectors representing the plurality of database tables in the multi-dimensional vector space in response to the indication, and grouping the plurality of database tables into one or more table clusters based on positions of the respective vectors representing the plurality of database tables in the multi-dimensional vector space.

Abstract: Various examples are directed to systems and methods for utilizing relationship data in a computing system. The computing system may extract first relationship data from a document and determine a first confidence value describing the first relationship data. The computing system may write the first relationship data to a knowledge graph data structure. The computing system may serve a first user interface page to a user computing device associated with a first user and receive feedback data describing an accuracy of the first relationship data. The computing system may modify a first confidence subunit of a triple data unit associated with the relationship to describe an updated confidence value based on the feedback data and a trust score of the first user.