Context-aware entity linking for knowledge graphs to support decision making

ChangeBridge: Patent Grants - AI & Computing (G06N)

Published April 7th, 2026

Detected April 7th, 2026

Summary

The USPTO granted NEC Corporation Patent No. US12596933B2 for a machine learning model combining context transformers with self-attention layers for knowledge graph link prediction. The model processes query embeddings (subject, object, relation) against knowledge graph data to generate attention scores and significance metrics. The decision head integrates these outputs to support AI-driven decision-making systems.

View original document View source feed page

What changed

NEC Corporation received a patent for a neural network system that performs context-aware entity linking within knowledge graphs. The system uses a context transformer with self-attention layers to process input embeddings including query sets (subject, object, relation) and knowledge graph embeddings. Each self-attention layer generates attention scores, while the final layer produces link predictions and outputs for each embedding. A decision head then combines attention scores with outputs to determine significance scores for inputs.

For technology companies and AI developers, this patent represents protected intellectual property in knowledge graph and decision-support AI systems. Organizations developing similar machine learning architectures for entity linking, graph-based reasoning, or decision-making applications should consider potential licensing requirements or design-around strategies to avoid infringement. Patent grants establish enforceable intellectual property rights but do not create compliance obligations for third parties.

What to do next

Monitor for updates

Source document (simplified)

← USPTO Patent Grants

Context-aware entity linking for knowledge graphs to support decision making

Grant US12596933B2 Kind: B2 Apr 07, 2026

Assignee

NEC CORPORATION

Inventors

David Friede, Kiril Gashteovski

Abstract

A machine learning model includes a context transformer and a decision head. The context transformer is a neural network of self-attention layers. The model makes a link prediction for a query embedding. Input embeddings are received at inputs of the context transformer. The input embeddings have: a query embedding set, the query embedding set comprising a subject embedding, object embedding, and relation embedding, one of the subject embedding, the object embedding, and the relation embedding being the query embedding; and knowledge graph embeddings. A first self-attention layer generates an attention score for each of the input embeddings. A final layer of the context transformer generates the link prediction for the query embedding and an output associated with each of the input embeddings. The decision head combines the attention score and the output for each of the input embeddings to determine a significance score for each of the input embeddings.