Symbolic Discovery of Optimization Algorithms

杠杆(统计) 计算机科学 算法 一般化 单调函数 深层神经网络 简单(哲学) 人工神经网络 人工智能 机器学习 数学 认识论 数学分析 哲学
作者
Xiangning Chen,Liang Chen,Da Huang,Esteban Real,Kaiyuan Wang,Yao Liu,Hieu Pham,Xuanyi Dong,Thang M. Luong,Cho‐Jui Hsieh,Yifeng Lu,Quoc V. Le
出处
期刊:Cornell University - arXiv [Cornell University]
被引量:92
标识
DOI:10.48550/arxiv.2302.06675
摘要

We present a method to formulate algorithm discovery as program search, and apply it to discover optimization algorithms for deep neural network training. We leverage efficient search techniques to explore an infinite and sparse program space. To bridge the large generalization gap between proxy and target tasks, we also introduce program selection and simplification strategies. Our method discovers a simple and effective optimization algorithm, $\textbf{Lion}$ ($\textit{Evo$\textbf{L}$ved S$\textbf{i}$gn M$\textbf{o}$me$\textbf{n}$tum}$). It is more memory-efficient than Adam as it only keeps track of the momentum. Different from adaptive optimizers, its update has the same magnitude for each parameter calculated through the sign operation. We compare Lion with widely used optimizers, such as Adam and Adafactor, for training a variety of models on different tasks. On image classification, Lion boosts the accuracy of ViT by up to 2% on ImageNet and saves up to 5x the pre-training compute on JFT. On vision-language contrastive learning, we achieve 88.3% $\textit{zero-shot}$ and 91.1% $\textit{fine-tuning}$ accuracy on ImageNet, surpassing the previous best results by 2% and 0.1%, respectively. On diffusion models, Lion outperforms Adam by achieving a better FID score and reducing the training compute by up to 2.3x. For autoregressive, masked language modeling, and fine-tuning, Lion exhibits a similar or better performance compared to Adam. Our analysis of Lion reveals that its performance gain grows with the training batch size. It also requires a smaller learning rate than Adam due to the larger norm of the update produced by the sign function. Additionally, we examine the limitations of Lion and identify scenarios where its improvements are small or not statistically significant. Lion is also successfully deployed in production systems such as Google search ads CTR model.

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
xxv发布了新的文献求助10
刚刚
colourful完成签到,获得积分10
刚刚
深情安青的应助被can采纳,获得10
1秒前
典雅绮露完成签到,获得积分10
1秒前
烂漫时完成签到,获得积分10
1秒前
mate发布了新的文献求助10
1秒前
顾矜的应助被抠抠小手采纳,获得10
2秒前
2秒前
2秒前
YSG完成签到 ,获得积分10
3秒前
852的应助被感动的亦云采纳,获得10
3秒前
3秒前
3秒前
春去春来完成签到,获得积分10
3秒前
3秒前
4秒前
4秒前
muky完成签到,获得积分10
4秒前
5秒前
曹文鹏发布了新的文献求助10
5秒前
LY发布了新的文献求助10
6秒前
徐开心完成签到,获得积分10
6秒前
6秒前
6秒前
Zylah发布了新的文献求助10
7秒前
不吃蛋黄完成签到,获得积分10
7秒前
7秒前
隐形曼青的应助被mate采纳,获得30
8秒前
辰辰发布了新的文献求助10
8秒前
木子发布了新的文献求助10
9秒前
水天需发布了新的文献求助10
9秒前
9秒前
10秒前
1122发布了新的文献求助10
10秒前
上官若男的应助被搞怪孤丝采纳,获得10
10秒前
科研通AI6.3的应助被牧百川采纳,获得10
11秒前
可爱的函函的应助被普鲁卡因采纳,获得10
11秒前
11秒前
hwj发布了新的文献求助10
11秒前
晓晓发布了新的文献求助10
11秒前
高分求助中
(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
拉伸高分子熔体中缠结行为的模拟和小角中子散射研究 1700
The World Atlas of Wine 9th edition 1000
奈特基础组织学:配套组织病理学(第4版)(Netter's Essential Histology With Correlated Histopathology) 800
八旗制度の研究 600
Improving the wear resistance of vehicle gear oil by oil soluble LaF3 nanoparticles and typical additives of gear oil 500
Boundaries and Bodhisattvas 500
热门求助领域 (近24小时)
化学 材料科学 医学 生物 纳米技术 计算机科学 工程类 有机化学 化学工程 生物化学 物理 复合材料 内科学 催化作用 光电子学 人工智能 细胞生物学 无机化学 物理化学 数学
热门帖子
关注 科研通微信公众号,转发送积分 8183894
求助须知:如何正确求助?哪些是违规求助? 9573067
关于积分的说明 21829226
捐赠科研通 7780983
什么是DOI,文献DOI怎么找? 3434150
关于科研通互助平台的介绍 2558315
邀请新用户注册赠送积分活动 2480648