International Business Machines Corporation

Joshua F Payne

Arjun Natarajan

Ashish Kundu

Kapil Kumar Singh

Roger C Raphael

Scott Schumacher

Mu Qiao

Mustafa Canim

Rajesh M Desai

Aniya Aggarwal

Sanjana Sahayaraj

Alexandre Rademaker

An apparatus and related method defend against adversarial queries. A policy enforcement hypergraph is constructed to express a set of security policies. Then, the hypergraph is repeatedly traversed to determine whether a user behavior is changing over time. The user behavior is measured by reference to a vertex or an edge in the hypergraph. If it is determined that the user behavior has changed over time an enforcement action is taken based on a security policy.

Defending against adversarial queries in a data governance system

A processor-implemented method securely responds to a query for information from a data graph. One or more processors create an embedding for encrypted sensitive information in vertices in a data graph; and bucketize embedded encrypted sensitive information on an embedding graph, where bucketizing the embedded encrypted sensitive information clusters vertices from the graph database that have shared data graph features. The processor(s) receive a query of the data graph from a requester, where the query is for information related to the shared data graph features. The processor(s) retrieve a bucket from the embedding graph that contains the information related to the shared data graph features; and extract encrypted sensitive information from the retrieved bucket.

Secure query processing on graph stores

A deep-learning based method evaluates similarities of entities in decentralized identity graphs. One or more processors represent a first identity profile as a first identity graph and a second identity profile as a second identity graph. The processor(s) compare the first identity graph to the second identity graph, which are decentralized identity graphs from different identity networks, in order to determine a similarity score between the first identity profile and the second identity profile. The processor(s) then implement a security action based on the similarity score.

Deep learning-based similarity evaluation in decentralized identity graphs

A method, computer system, and a computer program product for text bias identification and correction are provided. A first text corpus may be received. A designation of a second text corpus may be received. Words of the first text corpus may be embedded as a first word embedding in an embedding model. The first word embedding may be compared to a second word embedding in the embedding model to identify a first biased text in the first text corpus. The second word embedding may be from the second text corpus. A first replacement text portion may be generated as a substitute for the first biased text. The first replacement text portion may include a first unbiased text. The first biased text and the first replacement text portion may be presented.

Bias identification and correction in text documents

Preserving distributions of data values of a data asset in a data anonymization operation is provided. Anonymizing data values is performed by transforming sensitive data in a set of columns over rows of the data asset while preserving distribution of the data values in the set of transformed columns to a defined degree using a set of autoencoders and loss function. The autoencoders are base trained from preexisting data in a data assets catalog and actively trained during data dissemination. Parametric coefficients of the loss function are configured and the threshold is generated using policies from an enforcement decision for the data asset and data consumer. The loss function value of a selected row is compared to the threshold. Transformed data values of the selected row are transcribed to an output row when the loss function value is greater than the threshold and disseminated to the data consumer.

Learning to transform sensitive data with variable distribution preservation

A method creates an embedding for an unlabeled vertex in a hypergraph. The method includes receiving a hypergraph of hyperedges, where each of the hyperedges includes one or more vertices, and at least one of the hyperedges includes an unlabeled vertex; generating a hypergraph of vertices from the hypergraph of hyperedges, where each of the vertices in the hypergraph of vertices includes one or more of the one or more hyperedges from the hypergraph of hyperedges; performing a first type of random walk through the hypergraph of hyperedges; performing a second type of random walk through the hypergraph of vertices; generating a set of vertex embeddings from the first type of random walk and a set of hyperedge embeddings from the second type of random walk; and using results of the first and second random walks to train a neural network to create an embedding for the unlabeled vertex.

Training a neural network to create an embedding for an unlabeled vertex in a hypergraph

A method provides a security action based on identity profile scores. One or more processors represent an identity profile as a knowledge graph. The processor(s) associate a set of changes of the identity profile across a plurality of identity networks with a fraud score. The processor(s) then implement a security action based on the fraud score.

Deep learning-based identity fraud detection

A method provides a network-agnostic identity broker for retrieving identity records across heterogeneous identity networks. An identity broker receives a client request from a client to retrieve and evaluate user identity information for confirming an identity of a particular entity. The identity broker utilizes a group membership of the client to select a set of policies for handling the client request, and selects an identity network from multiple heterogeneous identity networks as a selected identity network to which the client request is to be sent. The identity broker sends the client request to the selected identity network, and then receives a response from the selected identity network. The identity broker evaluates the response according to the set of policies, such that the evaluated response conforms with the set of policies, and transmits the evaluated response to the client.

Adaptive identity broker for governance of decentralized identities across multiple heterogeneous identity networks

A method trains a neural network to recognize whether a resource is authorized to be returned to a requester. One or more processors train a neural network to traverse a policy enforcement hypergraph in order to identify a security policy to be used for a resource request and to authorize a use of a requested resource by a requester. The policy enforcement hypergraph is derived from a policy enforcement graph that expresses a set of security profiles for resources and requesters. The processor(s) receive a resource request for a requested resource from a requester, where the resource request includes a description of the requester. A system/user inputs a description of the received resource request and a description of the policy enforcement hypergraph into the trained neural network in order to selectively return the requested resource to the requester.

Resource access policy enforcement using a hypergraph

A method routes an identity query to a particular identity network. An identity broker determines that a candidate identity network is associated with a confidence score that satisfies predetermined criteria. Responsive to determining that the candidate identity network is associated with the confidence score that satisfies the predetermined criteria, the identity broker onboards the candidate identity network into a set of identity networks services, and then routes an identity query for an identity to the candidate identity network that satisfies the predetermined criteria.

Growing community of inventors

San Antonio, TX, United States of America

Joshua F Payne