Nvidia patents method to shrink neural networks by removing unused parts
Nvidia describes a technique to make artificial intelligence models smaller by identifying and removing specific components called filters. The system groups these filters across different layers of the network and uses a mathematical process to ensure that removing them does not break the model's structure.
In the filing’s own words
Systems and methods to use weight information corresponding to a neural network, wherein weight information of at least a first one or more layers of the neural network is modified to compensate for deactivating one or more second layers of the neural network. The neural network may be a residual neural network.
What exactly they claim (claim 1)
1. One or more processors, comprising: circuitry to: group one or more filters across two or more layers of a neural network; perform a regularization on the group of the one or more filters to enforce aligned structure sparsity across the two or more layers; and remove the group of the one or more filters from the two or more layers of the neural network based, at least in part, on a result of the regularization and a pruning criteria.
Facts
- Company
- Nvidia (NVIDIA Corporation)
- Official title
- VARIANCE AWARE TECHNIQUES TO PRUNE NEURAL NETWORKS
- Publication number
- US20260310965A1
- Application number
- 19710162
- Filed
- Jun 16, 2026
- Published
- Oct 8, 2026
- Inventors
- 3
- Claims
- 20
Classified under
G06N 3/082 Physics & computingG06N 3/04 Physics & computingG06V 30/18057 Physics & computingMore from Nvidia in AI & software
- Nvidia patents AI coach that analyzes your gameplay to suggest strategies Oct 13, 2026
- Nvidia patents AI system that writes and voices character conversations Sep 10, 2026
- Nvidia patents system that uses AI to find bugs in game videos Sep 17, 2026
- Nvidia patents AI method to make digital characters move realistically Sep 8, 2026
- Nvidia patents AI system that builds image generators from text prompts Sep 15, 2026
Summaries are simplified from the filing’s own abstract and first claim. A patent filing is not a product announcement.
Patent data and drawings: U.S. Patent and Trademark Office (USPTO). Nebula is not affiliated with or endorsed by the USPTO.