已入深夜,您辛苦了!由于当前在线用户较少,发布求助请尽量完整地填写文献信息,科研通机器人24小时在线,伴您度过漫漫科研夜!祝你早点完成任务,早点休息,好梦!

MoIL: Momentum Imitation Learning for Efficient Vision-Language Adaptation

计算机科学 人工智能 适应(眼睛) 模仿 计算机视觉 机器学习 心理学 光学 物理 社会心理学
作者
Gen Luo,Yiyi Zhou,Minglang Huang,Tianhe Ren,Xiaoshuai Sun,Rongrong Ji
出处
期刊:IEEE Transactions on Pattern Analysis and Machine Intelligence [IEEE Computer Society]
卷期号:: 1-13 被引量:1
标识
DOI:10.1109/tpami.2024.3435790
摘要

Pre-training and fine-tuning have been the de-facto paradigm in vision-language domains. Along with the rapid growth of model sizes, fully fine-tuning these large-scale vision-language pre-training (VLP) models requires prohibitively expensive storage costs. To address this issue, recent advances in NLP offer a promising and efficient adaptation approach called LoRA, which aims to approximate the fine-tuning of large pre-trained model by updating low-rank parameters. Despite its effectiveness, we identify that LoRA suffers a large approximation error on VLP models and its optimization is also inefficient, which greatly limits its performance upper bound. In this paper, we mathematically prove that the approximation error of low-rank adaptation can be optimized by a new optimization objective, i.e., the weight distance between LoRA and fine-tuning. Based on this finding, we propose a novel PETL method for VLP models, namely momentum imitation learning (MoIL). Specifically, MoIL formulates PETL as a weight imitation learning process and directly optimize the approximation error bound of the low-rank adaptation. Based on this training scheme, we also explore a new hybrid approximation function to reduce the learning difficulty of low-rank adaptations. With these two novel designs, MoIL can greatly improve the optimization efficiency of the low-rank parameters on VLP models. We validate MoIL on three VLP models ranging from end-to-end network to two-stage network, and conduct extensive experiments on four VL tasks. Experimental results demonstrate superior performance and optimization efficiency of MoIL than existing PETL methods. For instance, by updating only 6.23% parameters, MoIL can even outperform full tuning by +2.3% on image-text matching task. Meanwhile, its inference efficiency and generalization ability is also validated by multiple VLP models, e.g., VLMO and VinVL.
最长约 10秒,即可获得该文献文件

科研通智能强力驱动
Strongly Powered by AbleSci AI
科研通是完全免费的文献互助平台,具备全网最快的应助速度,最高的求助完成率。 对每一个文献求助,科研通都将尽心尽力,给求助人一个满意的交代。
实时播报
Owen应助停云采纳,获得10
2秒前
顾先森发布了新的文献求助10
3秒前
墨墨小7完成签到,获得积分10
3秒前
柳叶笺完成签到,获得积分10
3秒前
肉丸肉饼666完成签到,获得积分10
4秒前
Nokia应助诚心求文采纳,获得10
7秒前
科研通AI6.3应助柳叶笺采纳,获得10
8秒前
8秒前
8秒前
17秒前
courage完成签到,获得积分10
18秒前
诚心求文完成签到,获得积分10
21秒前
12等等发布了新的文献求助10
21秒前
24秒前
25秒前
xiaojunsong完成签到 ,获得积分10
25秒前
wanci应助YLC采纳,获得10
25秒前
25秒前
小蘑菇应助12等等采纳,获得10
26秒前
27秒前
勤劳惜梦发布了新的文献求助10
28秒前
ZJ发布了新的文献求助10
30秒前
shine发布了新的文献求助10
30秒前
脑洞疼应助寿司求学记采纳,获得10
30秒前
拾新完成签到,获得积分10
32秒前
32秒前
柳叶笺发布了新的文献求助10
34秒前
坚强煜城发布了新的文献求助10
34秒前
YLC发布了新的文献求助10
35秒前
36秒前
计划逃跑发布了新的文献求助10
39秒前
42秒前
无花果应助坚强煜城采纳,获得10
43秒前
fancymao发布了新的文献求助10
47秒前
冷静完成签到,获得积分10
47秒前
科研通AI6.3应助徐木木采纳,获得10
49秒前
49秒前
50秒前
54秒前
JamesPei应助DJ采纳,获得10
54秒前
高分求助中
Principles of Economics, 11th Edition 10000
University Physics with Modern Physics, 16th edition 10000
(应助此贴封号)【重要!!请各用户(尤其是新用户)详细阅读】【科研通的精品贴汇总】 10000
48V Low-voltage Power Distribution Network (PDN) Architecture Industry Report, 2024 800
ズームレンズの光学設計に関する研究 800
Fundamentals of Pharmaceutical and Biologics Regulations: A Global Perspective, Second Edition 700
Matrix Methods in Data Mining and Pattern Recognition Second Edition 610
热门求助领域 (近24小时)
化学 材料科学 医学 生物 纳米技术 工程类 有机化学 化学工程 生物化学 计算机科学 内科学 物理 复合材料 催化作用 细胞生物学 无机化学 光电子学 物理化学 电极 基因
热门帖子
关注 科研通微信公众号,转发送积分 7297191
求助须知:如何正确求助?哪些是违规求助? 8915665
关于积分的说明 18878769
捐赠科研通 6962972
什么是DOI,文献DOI怎么找? 3210516
关于科研通互助平台的介绍 2379824
邀请新用户注册赠送积分活动 2186984