USPTO Patent Grant: Heterogeneous Tree Graph Neural Network for Label Prediction

ChangeBridge: Patent Grants - Business Methods (G06Q)

Published March 24th, 2026

Detected March 25th, 2026

Summary

The USPTO has granted a patent to Microsoft Technology Licensing, LLC for a method using heterogeneous tree graph neural networks for label prediction. This method involves identifying node-target paths, assigning path type identifiers, and extracting a semantic tree to generate metapath embeddings for label prediction.

View original document View source feed page

What changed

The United States Patent and Trademark Office (USPTO) has granted patent US12585946B2 to Microsoft Technology Licensing, LLC. The patent covers a method for label prediction within a heterogeneous graph structure. Key aspects include identifying node-target paths, assigning path type identifiers based on edges and types, and extracting a semantic tree. This semantic tree is then encoded using neural networks to generate metapath embeddings, which are used to predict labels for target nodes within the graph.

This patent grant is primarily of informational value, detailing a novel technical approach to data analysis and prediction using graph neural networks. While it does not impose new regulatory obligations, companies in the technology sector, particularly those involved in AI, machine learning, and data analytics, may find the patented methodology relevant to their research and development efforts. No immediate compliance actions are required, but it signifies a specific area of innovation in business method patents.

Source document (simplified)

← USPTO Patent Grants

Heterogeneous tree graph neural network for label prediction

Grant US12585946B2 Kind: B2 Mar 24, 2026

Assignee

Microsoft Technology Licensing, LLC

Inventors

Qinlong Luo, Mingyu Guan, Jack Wilson Stokes, III, Purvanshi Mehta, Elnaz Nouri, Fuchen Liu

Abstract

A method for making predictions pertaining to entities represented within a heterogeneous graph includes: identifying, for each node in the heterogeneous graph structure, a set of node-target paths that connect the node to a target node; assigning, to each of the node-target paths identified for each node, a path type identifier indicative of a number of edges and corresponding edge types in the associated node-target path; and extracting a semantic tree from the heterogeneous graph structure. The semantic tree includes the target node as a root node and defines a hierarchy of metapaths that each individually correspond to a subset of the node-target paths in the heterogeneous graph structure assigned to a same path type identifier. The semantic tree is encoded, using one or more neural networks by generating a metapath embedding corresponding to each metapath in the semantic tree. Each of the resulting metapath embeddings encodes aggregated feature-label data for nodes in the heterogeneous graph structure corresponding to the path type identifier corresponding to the metapath associated with the metapath embedding. A label is predicted for the target node in the heterogeneous graph structure based on the set of metapath embeddings.