Abstract: Techniques for mapping human-understandable attributes to a neural network are disclosed. A system generates human-understandable explanations for decisions or predictions of a neural network by mapping human-understandable attributes to the trained neural network. The system may map a human-understandable attribute to a particular label, indicating a correlation between the particular human-understandable attribute and the particular label. In addition, or in the alternative, the system may map the human-understandable attribute to a particular layer of the neural network, such as a particular hidden layer, indicating a correlation between the human-understandable attribute and the particular layer.

Abstract: Embodiments of the present invention relate to techniques for creating policies. A plurality of objects representative of semantic objects are provided to a user. An arrangement of a subset of the objects, the arrangement representative of a policy, is received. The arrangement is converted to instructions for implementation by an application configured to implement policies. One or more of the objects may include fields and/or controls for specifying criteria of semantic objects represented by the objects.

Abstract: Embodiments of the present invention relate to techniques for creating policies. A plurality of objects representative of semantic objects are provided to a user. An arrangement of a subset of the objects, the arrangement representative of a policy, is received. The arrangement is converted to instructions for implementation by an application configured to implement policies. One or more of the objects may include fields and/or controls for specifying criteria of semantic objects represented by the objects.

Abstract: Embodiments of the present invention relate to techniques for creating policies. A plurality of objects representative of semantic objects are provided to a user. An arrangement of a subset of the objects, the arrangement representative of a policy, is received. The arrangement is converted to instructions for implementation by an application configured to implement policies. One or more of the objects may include fields and/or controls for specifying criteria of semantic objects represented by the objects.

Abstract: Résumé value extraction based on a résumé type is disclosed. An analysis model, for analyzing a résumé, is selected based on a résumé type of the résumé. An example of a résumé type is a job industry associated with the résumé. Different analysis models are used to analyze résumés corresponding to different résumé types. The analysis model is used to identify résumé values, in the résumé, that correspond to résumé attributes associated with the résumé type. Additionally or alternatively, the analysis model is used to identify segments, in the résumé, that correspond to segment types associated with the résumé type. After a segment type of a particular segment is identified, résumé values within the particular segment are analyzed based on the segment type of the particular segment.

Abstract: Embodiments of the present invention allow a system to use data gathered from social networks and other systems to determine an ordered list of desired topics or skills to define a subject matter area and incorporate the order of the topics or skills into a search. To define this subject matter area, embodiments can consider not just the topics or skills that are listed, but those topics or skills that are similar based on a pre-computed topic graph. These considerations can be incorporated into a generated query, so that the query itself accounts for similarity of topics via the topic graph and the order of desired terms. The query generation process can include a claimed skills veracity model that provides differential weighting to claimed skills, based on the skill-sets of users who are deemed to be similar to the user being evaluated.

Abstract: Embodiments provide for assessing and scoring user proficiency in topics determined by data from social networks and other sources. Embodiments can combine the information available to direct models, e.g., using user profiles, endorsements, etc., with data from various other external knowledge bases to find implicit topics for users, and other types of reasoning to compute proficiency scores. Implicit topics, i.e., those topics related to explicitly identified areas of expertise, can be determined based on a topic graph such as a Directed Acyclic Graph (DAG). Generally speaking, embodiments can traverse the topic graph for explicitly identified skills or topics and determine related or similar new skills based on nearby nodes of the graph. This approach can uncover skills that user that may not have disclosed as well as scoring users on skills based on the skill's similarity to those the user did claim.

Abstract: Résumé value extraction based on a résumé type is disclosed. An analysis model, for analyzing a résumé, is selected based on a résumé type of the résumé. An example of a résumé type is a job industry associated with the résumé. Different analysis models are used to analyze résumés corresponding to different résumé types. The analysis model is used to identify résumé values, in the résumé, that correspond to résumé attributes associated with the résumé type. Additionally or alternatively, the analysis model is used to identify segments, in the résumé, that correspond to segment types associated with the résumé type. After a segment type of a particular segment is identified, résumé values within the particular segment are analyzed based on the segment type of the particular segment.

Abstract: Embodiments of the present invention relate to techniques for creating policies. A plurality of objects representative of semantic objects are provided to a user. An arrangement of a subset of the objects, the arrangement representative of a policy, is received. The arrangement is converted to instructions for implementation by an application configured to implement policies. One or more of the objects may include fields and/or controls for specifying criteria of semantic objects represented by the objects.

Abstract: Embodiments of the present invention relate to techniques for creating policies. A plurality of objects representative of semantic objects are provided to a user. An arrangement of a subset of the objects, the arrangement representative of a policy, is received. The arrangement is converted to instructions for implementation by an application configured to implement policies. One or more of the objects may include fields and/or controls for specifying criteria of semantic objects represented by the objects.

Abstract: User profiles can be analyzed to identify profiles matching to the same identity. For example, profiles from different social network systems are analyzed to determine if the profiles are associated with the same user of the social network systems. Multiple heuristics may be calculated using different algorithms. The calculated heuristics may then be combined to generate a match score that indicates whether two profiles match.

Abstract: Techniques for enforcing policies. A set of data is stored in one or more data stores. A plurality of semantic concepts for an ontology are defined and a map is constructed from the set of data to the concepts. The map is executed in order to transform data from the set of data to a second set of data. The second set of data is stored according to the ontology. The second set of data is reasoned in order to determine compliance with one or more policies.

Abstract: Embodiments provide for assessing and scoring user proficiency in topics determined by data from social networks and other sources. Embodiments can combine the information available to direct models, e.g., using user profiles, endorsements, etc., with data from various other external knowledge bases to find implicit topics for users, and other types of reasoning to compute proficiency scores. Implicit topics, i.e., those topics related to explicitly identified areas of expertise, can be determined based on a topic graph such as a Directed Acyclic Graph (DAG). Generally speaking, embodiments can traverse the topic graph for explicitly identified skills or topics and determine related or similar new skills based on nearby nodes of the graph. This approach can uncover skills that user that may not have disclosed as well as scoring users on skills based on the skill's similarity to those the user did claim.

Abstract: Techniques for displaying information. Policy violations are identified, based at least in part on data stored in a data store. For the policy violations, a plurality of semantic objects related to the violations are identified. Arrangements of graphical objects are displayed where the graphical objects represent the identified semantic objects and where the arrangement indicates one or more relationships between pairs of the semantic objects.

Abstract: Techniques for implementing policies. In an embodiment, first data is stored in a first data store according to a first schema. A second schema is defined based at least in part on a policy and an ontology. Second data, which includes at least a portion of the first data, is stored in a second data store according to the second schema. Storing the second data is based at least in part on a mapping of the first schema to the second schema. At least a portion of the second data is analyzed and results of the analysis are provided to a user.

Abstract: Embodiments of the present invention allow a system to use data gathered from social networks and other systems to determine an ordered list of desired topics or skills to define a subject matter area and incorporate the order of the topics or skills into a search. To define this subject matter area, embodiments can consider not just the topics or skills that are listed, but those topics or skills that are similar based on a pre-computed topic graph. These considerations can be incorporated into a generated query, so that the query itself accounts for similarity of topics via the topic graph and the order of desired terms. The query generation process can include a claimed skills veracity model that provides differential weighting to claimed skills, based on the skill-sets of users who are deemed to be similar to the user being evaluated.

Abstract: Techniques for displaying information. Policy violations are identified, based at least in part on data stored in a data store. For the policy violations, a plurality of semantic objects related to the violations are identified. Arrangements of graphical objects are displayed where the graphical objects represent the identified semantic objects and where the arrangement indicates one or more relationships between pairs of the semantic objects.

Abstract: Reputation metrics are used to gauge risk of individuals to an organization, such as employees of a business. The reputation metrics may be calculated from both internal and external data sources, including social network profiles of the individuals. Calculations of risk are used to make determinations regarding the activities the individuals are authorized to engage in.

Abstract: Techniques, including systems and methods, for generating data are disclosed and suggested herein. Original data used in connection with one or more applications is analyzed in order to determine one or more distribution characteristics for the original data. The distribution characteristics are used to generate data that is similarly distributed. The generated data may be used as seed data for demonstrating, testing, or otherwise using one or more applications.

Abstract: Techniques for reasoning data are disclosed. The techniques may include storing data in an ontology. The data may be partitioned and/or distributed to a plurality of reasoners. Processed data from the plurality of reasoners may then be combined, the processed data having been received from the plurality of reasoners and/or having been processed by the plurality of reasoners.