Alert button
Picture for Zhenmei Shi

Zhenmei Shi

Alert button

Conv-Basis: A New Paradigm for Efficient Attention Inference and Gradient Computation in Transformers

Add code
Bookmark button
Alert button
May 08, 2024
Jiuxiang Gu, Yingyu Liang, Heshan Liu, Zhenmei Shi, Zhao Song, Junze Yin

Figure 1 for Conv-Basis: A New Paradigm for Efficient Attention Inference and Gradient Computation in Transformers
Figure 2 for Conv-Basis: A New Paradigm for Efficient Attention Inference and Gradient Computation in Transformers
Figure 3 for Conv-Basis: A New Paradigm for Efficient Attention Inference and Gradient Computation in Transformers
Viaarxiv icon

Exploring the Frontiers of Softmax: Provable Optimization, Applications in Diffusion Model, and Beyond

Add code
Bookmark button
Alert button
May 06, 2024
Jiuxiang Gu, Chenyang Li, Yingyu Liang, Zhenmei Shi, Zhao Song

Viaarxiv icon

Towards Few-Shot Adaptation of Foundation Models via Multitask Finetuning

Add code
Bookmark button
Alert button
Feb 22, 2024
Zhuoyan Xu, Zhenmei Shi, Junyi Wei, Fangzhou Mu, Yin Li, Yingyu Liang

Viaarxiv icon

Fourier Circuits in Neural Networks: Unlocking the Potential of Large Language Models in Mathematical Reasoning and Modular Arithmetic

Add code
Bookmark button
Alert button
Feb 12, 2024
Jiuxiang Gu, Chenyang Li, Yingyu Liang, Zhenmei Shi, Zhao Song, Tianyi Zhou

Viaarxiv icon

A Graph-Theoretic Framework for Understanding Open-World Semi-Supervised Learning

Add code
Bookmark button
Alert button
Nov 06, 2023
Yiyou Sun, Zhenmei Shi, Yixuan Li

Viaarxiv icon

Provable Guarantees for Neural Networks via Gradient Feature Learning

Add code
Bookmark button
Alert button
Oct 19, 2023
Zhenmei Shi, Junyi Wei, Yingyu Liang

Viaarxiv icon

When and How Does Known Class Help Discover Unknown Ones? Provable Understanding Through Spectral Analysis

Add code
Bookmark button
Alert button
Aug 09, 2023
Yiyou Sun, Zhenmei Shi, Yingyu Liang, Yixuan Li

Figure 1 for When and How Does Known Class Help Discover Unknown Ones? Provable Understanding Through Spectral Analysis
Figure 2 for When and How Does Known Class Help Discover Unknown Ones? Provable Understanding Through Spectral Analysis
Figure 3 for When and How Does Known Class Help Discover Unknown Ones? Provable Understanding Through Spectral Analysis
Figure 4 for When and How Does Known Class Help Discover Unknown Ones? Provable Understanding Through Spectral Analysis
Viaarxiv icon

Domain Generalization via Nuclear Norm Regularization

Add code
Bookmark button
Alert button
Mar 13, 2023
Zhenmei Shi, Yifei Ming, Ying Fan, Frederic Sala, Yingyu Liang

Figure 1 for Domain Generalization via Nuclear Norm Regularization
Figure 2 for Domain Generalization via Nuclear Norm Regularization
Figure 3 for Domain Generalization via Nuclear Norm Regularization
Viaarxiv icon

The Trade-off between Universality and Label Efficiency of Representations from Contrastive Learning

Add code
Bookmark button
Alert button
Feb 28, 2023
Zhenmei Shi, Jiefeng Chen, Kunyang Li, Jayaram Raghuram, Xi Wu, Yingyu Liang, Somesh Jha

Figure 1 for The Trade-off between Universality and Label Efficiency of Representations from Contrastive Learning
Figure 2 for The Trade-off between Universality and Label Efficiency of Representations from Contrastive Learning
Figure 3 for The Trade-off between Universality and Label Efficiency of Representations from Contrastive Learning
Figure 4 for The Trade-off between Universality and Label Efficiency of Representations from Contrastive Learning
Viaarxiv icon

A Theoretical Analysis on Feature Learning in Neural Networks: Emergence from Inputs and Advantage over Fixed Features

Add code
Bookmark button
Alert button
Jun 03, 2022
Zhenmei Shi, Junyi Wei, Yingyu Liang

Viaarxiv icon