Apparatus and methods for generating obfuscated data within a computing environment

Inventors

Aravamudan, Murali • Rajasekharan, Ajit

Assignees

Nference Inc

Abstract

An apparatus for generating obfuscated data within a computing environment, comprising a processor and a memory containing instructions configuring the processor to access a database containing a plurality of private data elements belonging to at least a private record, generate a set of obfuscated data elements, representative of the at least a private record, as a function of the plurality of private data elements using an generative model, determine a first distance measure between at least an obfuscated data element within the set of obfuscated data elements and at least a private data element of the plurality of private data elements within the database, and verify the first distance measure is within a distance range, wherein a minimum threshold of the distance range is determined as a function of a deidentification parameter and a maximum threshold of the distance range is determined as a function of an obfuscation parameter.

Core Innovation

The patent describes an apparatus for generating obfuscated data elements within a computing environment from private data elements belonging to a private record. The apparatus accesses a database containing a plurality of private data elements and generates, using a generative model, a set of obfuscated data elements representative of the at least a private record as a function of the plurality of private data elements.

After generating the set of obfuscated data elements, the apparatus determines a first distance measure between at least an obfuscated data element and at least a private data element of the plurality of private data elements within the database. The apparatus verifies that, for the at least an obfuscated data element within the set, the first distance measure is within a distance range, where the minimum threshold of the distance range is determined as a function of a deidentification parameter and the maximum threshold is determined as a function of an obfuscation parameter.

The apparatus further clusters the plurality of private data elements into a plurality of clusters using at least a clustering algorithm and selects a subset of obfuscated data elements from the set as a function of both the plurality of clusters and the verification. The selected subset of obfuscated data elements is transmitted to one or more requesting entities within a computing environment.

The patent also describes a corresponding method that performs the same sequence of operations, and related dependents that specify cluster-level verification based on a cluster centroid and a radius of influence, including determining a second distance measure and verifying that it lies within a specified distance range. The document further identifies example generative-model types and selection/scalability mechanisms, including particular generative model implementations such as an LLM and a transformer-based diffusion model.

Claims Coverage

The partial content provides two independent claims, covering both an apparatus and a method for generating and verifying obfuscated data elements. The independent claims share a common inventive structure: generative-model-based obfuscated data generation followed by distance-range verification parameterized by deidentification and obfuscation parameters, then clustering-based subset selection and transmission to requesting entities.

Across both independent claims, obfuscated data elements are generated via a generative model from private data elements, verified using a first distance measure constrained by a distance range defined by deidentification and obfuscation parameters, and then reduced to a transmitted subset via clustering-based selection that depends on both clustering and verification.

Stated Advantages

Documented Applications