Machine Learning Model Pruning System
Summary
The USPTO published patent application US20260094048A1 for a machine learning model pruning system developed by Amazon Technologies, Inc. The invention describes methods for optimizing neural network weights by strategically pruning connections and minimizing loss functions through batch processing techniques. This is a publication of a patent application filed on September 27, 2024.
What changed
The USPTO published Amazon Technologies' patent application for a machine learning model pruning system (Application No. 18899369, published April 2, 2026). The system performs iterative pruning passes where weights are set to zero, with additional passes conducted in batches containing both remaining and previously pruned weights. The innovation determines pruning adjustments by solving an optimization problem that minimizes loss functions, allowing for restoration of previously pruned weights or pruning of remaining weights.
Technology companies and researchers developing machine learning models should review this patent to assess potential licensing implications or freedom-to-operate considerations. While patent applications do not create immediate compliance obligations, understanding the technical scope of this pruning methodology may inform R&D strategy and model optimization approaches.
Archived snapshot
Apr 3, 2026GovPing captured this document from the original source. If the source has since changed or been removed, this is the text as it existed at that time.
MACHINE LEARNING MODEL PRUNING SYSTEM
Application US20260094048A1 Kind: A1 Apr 02, 2026
Assignee
Amazon Technologies, Inc.
Inventors
Gilad Amir Rosenberg, John Kyle Brubaker, Martin Schuetz, Helmut Gottfried Katzgraber
Abstract
Methods and apparatus for pruning weights of a trained machine learning model and making pruning adjustments to substantially optimize a loss function. In some embodiments, a machine learning model pruning system is configured to perform a first pruning pass of a machine learning model wherein at least a portion of the weights are set to zero. In some embodiments one or more additional pruning passes of the machine learning model may be performed in batches wherein each batch comprises one or more remaining weights and one or more previously pruned weights. In some embodiments, one or more pruning adjustments may be determined based on an optimization problem that minimizes a loss function for a given batch. In some embodiments, the pruning adjustment comprises restoring a previously pruned weight or pruning a remaining weight of the model.
CPC Classifications
G06N 20/00
Filing Date
2024-09-27
Application No.
18899369
Named provisions
Related changes
Get daily alerts for USPTO Patent Applications - AI & Computing (G06N)
Daily digest delivered to your inbox.
Free. Unsubscribe anytime.
Source
About this page
Every important government, regulator, and court update from around the world. One place. Real-time. Free. Our mission
Source document text, dates, docket IDs, and authority are extracted directly from USPTO.
The summary, classification, recommended actions, deadlines, and penalty information are AI-generated from the original text and may contain errors. Always verify against the source document.
Classification
Who this affects
Taxonomy
Browse Categories
Get alerts for this source
We'll email you when USPTO Patent Applications - AI & Computing (G06N) publishes new changes.
Subscribed!
Optional. Filters your digest to exactly the updates that matter to you.