GAN-based Neural Architecture Search Patent Grant

ChangeBridge: Patent Grants - AI & Computing (G06N)

Published April 7th, 2026

Detected April 7th, 2026

Summary

The USPTO granted Patent US12596914B2 to Huawei Technologies Co., Ltd. covering a method for neural architecture search using a generative adversarial network. The system comprises a generator that creates neural network architectures and a discriminator that selects the optimal architecture based on a user-provided search space query. The patent contains 20 claims and was filed on November 1, 2021.

View original document View source feed page

What changed

USPTO issued Patent US12596914B2 to Huawei Technologies Co., Ltd. for a GAN-based neural architecture search method and system. The patent describes a process where a generator creates multiple neural network architectures from a search space query, and a discriminator evaluates and selects the optimal architecture for transmission to the user device.

For technology companies developing AI and machine learning tools, this patent may create licensing considerations or potential freedom-to-operate concerns for similar neural architecture search implementations. Companies utilizing or commercializing GAN-based neural architecture search should review the patent claims to assess potential infringement exposure.

What to do next

Monitor for updates

Source document (simplified)

← USPTO Patent Grants

Generative adversarial neural architecture search

Grant US12596914B2 Kind: B2 Apr 07, 2026

Assignee

Huawei Technologies Co., Ltd.

Inventors

Seyed Saeed Changiz Rezaei, Fred Xuefei Han, Di Niu

Abstract

A method and system for neural architectural search (NAS) for performing a task. A generative adversarial network comprising a generator and a discriminator receives, from a user device, a query for neural network architecture, the query including a search space. The generator of the generative adversarial network generates a plurality of generated neural network architectures responsive to the received search space. The discriminator of the generative adversarial network selects an optimal neural network architecture from among the plurality of generated neural network architectures. The optimal generated neural network architecture is transmitted to the user device.