Skip to main navigation Skip to search Skip to main content

Research on personalized compression algorithm for pre-trained models based on homomorphic entropy increase

  • Xing Guo*
  • , Yicong Li
  • , Haohua Du
  • *Corresponding author for this work
  • Anhui University

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

This paper investigates the deployment challenges of Vision Transformer (ViT) and Large Language Models (LLMs). Vision Transformer captures global information through multi-head attention mechanisms, but its high computational cost limits its application on mobile devices. Although LLMs have achieved breakthroughs in natural language processing, they also face significant deployment challenges. To address these issues, we propose a hierarchical pruning strategy that distinguishes personalized layers from shared layers through compressed sensing and random sampling, significantly reducing model parameters. Experiments show that the hierarchical mechanism effectively balances pruning and accuracy, providing a new direction for deploying efficient and personalized AI models on mobile devices.

Original languageEnglish
Title of host publicationSecond International Conference on Optical Communication and Optoelectronic Technology, OCOT 2025
EditorsKlimis Ntalianis
PublisherSPIE
ISBN (Electronic)9781510699168
DOIs
StatePublished - 13 Nov 2025
Event2nd International Conference on Optical Communication and Optoelectronic Technology, OCOT 2025 - Changsha, China
Duration: 18 Jul 202520 Jul 2025

Publication series

NameProceedings of SPIE - The International Society for Optical Engineering
Volume13985
ISSN (Print)0277-786X
ISSN (Electronic)1996-756X

Conference

Conference2nd International Conference on Optical Communication and Optoelectronic Technology, OCOT 2025
Country/TerritoryChina
CityChangsha
Period18/07/2520/07/25

Keywords

  • hierarchical mechanism
  • large language models
  • mobile devices
  • model pruning
  • personalized AI
  • Vision transformer

Fingerprint

Dive into the research topics of 'Research on personalized compression algorithm for pre-trained models based on homomorphic entropy increase'. Together they form a unique fingerprint.

Cite this