Method and Apparatus for Lightweighting AI Model Using Knowledge Distillation and Pruning
Summary
The USPTO published patent application US20260099770A1 titled 'Method and Apparatus for Lightweighting AI Model Using Knowledge Distillation and Pruning.' The application (filed August 14, 2025) describes a method for reducing the computational complexity of AI models through knowledge distillation from a teacher model to a student model combined with pruning techniques. Inventors include Jae Ho Kim, Dong Hoon Lee, Se Jung Kim, and Yong Hyun Kwon. The CPC classification is G06N 20/00 (Machine Learning).
What changed
The USPTO published patent application US20260099770A1 covering methods and apparatus for lightweighting AI models using knowledge distillation and pruning. The application calculates loss values using outputs and feature vectors from teacher and student models to train and prune the student model for reduced computational complexity.
For AI developers and technology companies, this patent represents potential prior art for any AI optimization techniques involving knowledge distillation combined with pruning. Parties developing similar technologies should consider filing provisional applications or requesting file wrapper inspections to assess freedom-to-operate implications.
What to do next
- Monitor for patent grant or office action updates
Archived snapshot
Apr 11, 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.
METHOD AND APPARATUS FOR LIGHTWEIGHTING AI MODEL USING KNOWLEDGE DISTILLATION AND PRUNING
Application US20260099770A1 Kind: A1 Apr 09, 2026
Inventors
Jae Ho KIM, Dong Hoon LEE, Se Jung KIM, Yong Hyun KWON
Abstract
A method of lightweighting an AI model using knowledge distillation and pruning includes calculating a first loss value using an output value of a teacher model, an output value of a student model, a feature vector generated from the teacher model, a feature vector generated from the student model, and a ground truth, performing training on the student model so that the first loss value is minimized, and performing pruning on the student model using a second loss value calculated based on the feature vector generated from the teacher model and the feature vector generated from the student model.
CPC Classifications
G06N 20/00
Filing Date
2025-08-14
Application No.
19299975
Related changes
Get daily alerts for EPO Patent Bulletin - Pharma (A61K)
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 EPO Patent Bulletin - Pharma (A61K) publishes new changes.
Subscribed!
Optional. Filters your digest to exactly the updates that matter to you.