Yuanzhi Li

发表

Algorithmic Regularization in Over-parameterized Matrix Sensing and Neural Networks with Quadratic Activations

Hongyang Zhang, Yuanzhi Li, Tengyu Ma, 2017, COLT.

Algorithmic Regularization in Over-parameterized Matrix Recovery pdf

Yuanzhi Li, Tengyu Ma, Hongyang Zhang, 2017, ArXiv.

Algorithmic Framework for Model-based Deep Reinforcement Learning with Theoretical Guarantees pdf

Yuandong Tian, Trevor Darrell, Yuanzhi Li, 2018, ICLR.

RAND-WALK: A Latent Variable Model Approach to Word Embeddings

Sanjeev Arora, Yuanzhi Li, Yingyu Liang, 2015 .

Random Walks on Context Spaces: Towards an Explanation of the Mysteries of Semantic Word Embeddings pdf

Sanjeev Arora, Yuanzhi Li, Yingyu Liang, 2015, ArXiv.

Convergence Analysis of Two-layer Neural Networks with ReLU Activation

Yuanzhi Li, Yang Yuan, Yang Yuan, 2017, NIPS.

A Theoretical Analysis of NDCG Ranking Measures

Yuanzhi Li, Yining Wang, Liwei Wang, 2013 .

Recovery guarantee of weighted low-rank approximation via alternating minimization

Yuanzhi Li, Yingyu Liang, Andrej Risteski, 2016, ICML.

A Convergence Theory for Deep Learning via Over-Parameterization pdf

Yuanzhi Li, Zeyuan Allen-Zhu, Zhao Song, 2018, ICML.

Faster Principal Component Regression via Optimal Polynomial Approximation to sgn(x) pdf

Yuanzhi Li, Zeyuan Allen Zhu, Yuanzhi Li, 2016, ArXiv.

Learning and Generalization in Overparameterized Neural Networks, Going Beyond Two Layers

Yuanzhi Li, Yingyu Liang, Zeyuan Allen-Zhu, 2018, NeurIPS.

Much Faster Algorithms for Matrix Scaling

Avi Wigderson, Yuanzhi Li, Zeyuan Allen-Zhu, 2017, 2017 IEEE 58th Annual Symposium on Foundations of Computer Science (FOCS).

Random walks on discourse spaces: a new generative language model with applications to semantic word embeddings

Sanjeev Arora, Yingyu Liang, Tengyu Ma, 2015 .

Statistical convergence of the EM algorithm on Gaussian mixture models

Yuekai Sun, Ruofei Zhao, Yuanzhi Li, 2018, Electronic Journal of Statistics.

Feature Purification: How Adversarial Training Performs Robust Deep Learning

Yuanzhi Li, Zeyuan Allen-Zhu, 2020, 2021 IEEE 62nd Annual Symposium on Foundations of Computer Science (FOCS).

What Can ResNet Learn Efficiently, Going Beyond Kernels?

Yuanzhi Li, Zeyuan Allen-Zhu, 2019, NeurIPS.

L G ] 9 F eb 2 01 8 Make the Minority Great Again : First-Order Regret Bound for Contextual Bandits Zeyuan

Sébastien Bubeck, Yuanzhi Li, Zeyuan Allen-Zhu, 2018 .

AT THE E DGE OF S TABILITY

Ameet Talwalkar, J. Zico Kolter, J. Z. Kolter, 2021 .

ON THE ABILITY OF GRADIENT DESCENT TO LEARN NEURAL NETWORKS

Yuanzhi Li, Yuanzhi Li, 2018 .

Online Improper Learning in Games with Approximation Oracle

Yuanzhi Li, Wei Hu, Zhiyuan Li, 2018 .

Near Optimal Methods for Minimizing Convex Functions with Lipschitz $p$-th Derivatives

Yin Tat Lee, Yuanzhi Li, Qijia Jiang, 2019, COLT.

Near-Optimal Discrete Optimization for Experimental Design

Yuanzhi Li, Yining Wang, Aarti Singh, 2018 .

What Can ResNet Learn Efficiently, Going Beyond Kernels?

Yuanzhi Li, Zeyuan Allen-Zhu, Yuanzhi Li, 2019, NeurIPS.

When Is Generalizable Reinforcement Learning Tractable? pdf

Dhruv Malik, Pradeep Ravikumar, Yuanzhi Li, 2021, NeurIPS.

Learning (Very) Simple Generative Models Is Hard

Yuanzhi Li, Jungshian Li, Sitan Chen, 2022, NeurIPS.

Towards understanding how momentum improves generalization in deep learning

Yuanzhi Li, Samy Jelassi, 2022, ICML.

Fooling space-bounded models and low degree polynomials a survey

Yuanzhi Li, Xin Yang, 2013 .

Sampling is as easy as learning the score: theory for diffusion models with minimal data assumptions

Anru R. Zhang, Yuanzhi Li, Jungshian Li, 2022, ArXiv.

Dissecting adaptive methods in GANs

Yuanzhi Li, Gauthier Gidel, A. Mensch, 2022, ArXiv.

Feature Purification: How Adversarial Training Performs Robust Deep Learning

Yuanzhi Li, Zeyuan Allen-Zhu, 2020, 2021 IEEE 62nd Annual Symposium on Foundations of Computer Science (FOCS).

Near-optimal discrete optimization for experimental design: a regret minimization approach

Yuanzhi Li, Yining Wang, Aarti Singh, 2020, Mathematical Programming.

Toward Understanding the Feature Learning Process of Self-supervised Contrastive Learning

Yuanzhi Li, Zi-Hang Wen, 2021, ICML.

Linear Algebraic Structure of Word Senses, with Applications to Polysemy

Sanjeev Arora, Yuanzhi Li, Yingyu Liang, 2016, TACL.

Recovery Guarantee of Non-negative Matrix Factorization via Alternating Updates

Yuanzhi Li, Yingyu Liang, Andrej Risteski, 2016, NIPS.

On the Convergence Rate of Training Recurrent Neural Networks

Yuanzhi Li, Zeyuan Allen-Zhu, Zhao Song, 2018, NeurIPS.

A Latent Variable Model Approach to PMI-based Word Embeddings

Sanjeev Arora, Yuanzhi Li, Yingyu Liang, 2015, TACL.

Near-optimal discrete optimization for experimental design: a regret minimization approach

Yuanzhi Li, Aarti Singh, Yining Wang, 2017, Mathematical Programming.

Near-Optimal Design of Experiments via Regret Minimization

Yuanzhi Li, Yining Wang, Zeyuan Allen-Zhu, 2017, ICML.

Near-optimal method for highly smooth convex optimization

Yin Tat Lee, Yuanzhi Li, Qijia Jiang, 2018, COLT.

Doubly Accelerated Methods for Faster CCA and Generalized Eigendecomposition pdf

Yuanzhi Li, Zeyuan Allen-Zhu, Yuanzhi Li, 2016, ICML.

An Alternative View: When Does SGD Escape Local Minima? pdf

Yuanzhi Li, Yang Yuan, Robert D. Kleinberg, 2018, ICML.

Neon2: Finding Local Minima via First-Order Oracles

Yuanzhi Li, Zeyuan Allen-Zhu, Yuanzhi Li, 2017, NeurIPS.

Gradient Descent on Neural Networks Typically Occurs at the Edge of Stability pdf

Ameet Talwalkar, J. Zico Kolter, J. Z. Kolter, 2021, ICLR.

Learning Overparameterized Neural Networks via Stochastic Gradient Descent on Structured Data

Yuanzhi Li, Yingyu Liang, Yuanzhi Li, 2018, NeurIPS.

A Nearly Instance Optimal Algorithm for Top-k Ranking under the Multinomial Logit Model pdf

Xi Chen, Yuanzhi Li, Jieming Mao, 2017, SODA.

PET: Optimizing Tensor Programs with Partially Equivalent Transformations and Automated Corrections

Jidong Zhai, Shizhi Tang, Zhihao Jia, 2021, OSDI.

Geodesic cycles in random graphs

Yuanzhi Li, Lingsheng Shi, Yuanzhi Li, 2018, Discret. Math..

An optimal algorithm for bandit convex optimization pdf

Yuanzhi Li, Elad Hazan, Elad Hazan, 2016, ArXiv.

LoRA: Low-Rank Adaptation of Large Language Models pdf

Yelong Shen, Zeyuan Allen-Zhu, Phillip Wallis, 2021, ICLR.

Operator scaling via geodesically convex optimization, invariant theory and polynomial identity testing

Avi Wigderson, Yuanzhi Li, Ankit Garg, 2018, STOC.

Learning Over-Parametrized Two-Layer Neural Networks beyond NTK

Tengyu Ma, Yuanzhi Li, Hongyang R. Zhang, 2020, COLT.

Understanding the Generalization of Adam in Learning Neural Networks with Proper Regularization pdf

Quanquan Gu, Difan Zou, Yuan Cao, 2021, ICLR.

Sample Efficient Reinforcement Learning In Continuous State Spaces: A Perspective Beyond Linearity

Dhruv Malik, Vishwak Srinivasan, Aldo Pacchiano, 2021, ICML.

Local Signal Adaptivity: Provable Feature Learning in Neural Networks Beyond Kernels

Yuanzhi Li, Aarti Singh, Stefani Karp, 2021, NeurIPS.

Backward Feature Correction: How Deep Learning Performs Deep Learning

Yuanzhi Li, Zeyuan Allen-Zhu, 2020, ArXiv.

A law of robustness for two-layers neural networks

Sébastien Bubeck, Yuanzhi Li, Dheeraj M. Nagaraj, 2020, COLT.

Sampling is as easy as learning the score: theory for diffusion models with minimal data assumptions

Anru R. Zhang, Yuanzhi Li, Jungshian Li, 2022, ICLR.

The Mechanism of Prediction Head in Non-contrastive Self-supervised Learning

Yuanzhi Li, Zi-Hang Wen, Zixin Wen, 2022, NeurIPS.

Forward Super-Resolution: How Can GANs Learn Hierarchical Generative Models for Real-World Distributions pdf

Zeyuan Allen-Zhu, Yuanzhi Li, Yuanzhi Li, 2021, ArXiv.

Towards Explaining the Regularization Effect of Initial Large Learning Rate in Training Neural Networks

Colin Wei, Yuanzhi Li, Tengyu Ma, 2019, NeurIPS.

Linear Convergence of a Frank-Wolfe Type Algorithm over Trace-Norm Balls

Yuanzhi Li, Elad Hazan, Zeyuan Allen-Zhu, 2017, NIPS.

Chasing Nested Convex Bodies Nearly Optimally

Yin Tat Lee, Yuanzhi Li, Sébastien Bubeck, 2018, SODA.

Towards Understanding Ensemble, Knowledge Distillation and Self-Distillation in Deep Learning

Yuanzhi Li, Zeyuan Allen-Zhu, 2020, ICLR.

On the One-sided Convergence of Adam-type Algorithms in Non-convex Non-concave Min-max Optimization pdf

Zehao Dou, Yuanzhi Li, Yuanzhi Li, 2021, ArXiv.

Weighted Tallying Bandits: Overcoming Intractability via Repeated Exposure Optimality

Yuanzhi Li, Aarti Singh, Dhruv Malik, 2023, ICML.

Complete Policy Regret Bounds for Tallying Bandits

Yuanzhi Li, Yuanzhi Li, Dhruv Malik, 2022, COLT.

Follow the Compressed Leader: Faster Online Learning of Eigenvectors and Faster MMWU

Yuanzhi Li, Zeyuan Allen-Zhu, Yuanzhi Li, 2017, ICML.

Complexity of Highly Parallel Non-Smooth Convex Optimization

Yin Tat Lee, Yuanzhi Li, Qijia Jiang, 2019, NeurIPS.

Why Clean Generalization and Robust Overfitting Both Happen in Adversarial Training

Yuanzhi Li, Binghui Li, 2023, ArXiv.

Learning Polynomial Transformations via Generalized Tensor Decompositions

Anru R. Zhang, Yuanzhi Li, Jungshian Li, 2023, STOC.

First Efficient Convergence for Streaming k-PCA: A Global, Gap-Free, and Near-Optimal Rate

Yuanzhi Li, Zeyuan Allen-Zhu, Yuanzhi Li, 2016, 2017 IEEE 58th Annual Symposium on Foundations of Computer Science (FOCS).

Settling the Horizon-Dependence of Sample Complexity in Reinforcement Learning pdf

Yuanzhi Li, Lin F. Yang, Ruosong Wang, 2021, 2021 IEEE 62nd Annual Symposium on Foundations of Computer Science (FOCS).

A Theoretical Analysis of NDCG Type Ranking Measures pdf

Tie-Yan Liu, Liwei Wang, Yuanzhi Li, 2013, COLT.

The probability flow ODE is provably fast

Yuanzhi Li, Holden Lee, Jianfeng Lu, 2023, NeurIPS.

Fast Global Convergence of Online PCA pdf

Yuanzhi Li, Zeyuan Allen Zhu, Yuanzhi Li, 2016, ArXiv.

A Theoretical Analysis of Normalized Discounted Cumulative Gain (NDCG) Ranking Measures

Tie-Yan Liu, Yuanzhi Li, Yining Wang, 2013 .

Minimax Optimality (Probably) Doesn't Imply Distribution Learning for GANs pdf

Raghu Meka, Sitan Chen, Jerry Li, 2022, ICLR.

When can Wasserstein GANs minimize Wasserstein Distance?

Yuanzhi Li, Zehao Dou, 2020, ArXiv.

Making Method of Moments Great Again? -- How can GANs learn the target distribution

Yuanzhi Li, Zehao Dou, 2020 .

Making Method of Moments Great Again? -- How can GANs learn distributions

Yuanzhi Li, Zehao Dou, 2020 .

Even Faster SVD Decomposition Yet Without Agonizing Pain

Yuanzhi Li, Zeyuan Allen Zhu, Yuanzhi Li, 2016, NIPS.

Learning Over-Parametrized Two-Layer ReLU Neural Networks beyond NTK pdf

Yuanzhi Li, Tengyu Ma, Hongyang R. Zhang, 2020, COLT 2020.

2 Positive Result : The Learnability of Three-Layer ResNet W

Yuanzhi Li, Zeyuan Allen-Zhu, 2019 .

Learning Mixtures of Linear Regressions with Nearly Optimal Complexity

Yuanzhi Li, Yingyu Liang, Yuanzhi Li, 2018, COLT.

How Does Adaptive Optimization Impact Local Neural Network Geometry?

Yuanzhi Li, Dhruv Malik, Kaiqi Jiang, 2022, ArXiv.

Make the Minority Great Again: First-Order Regret Bound for Contextual Bandits pdf

Yuanzhi Li, Sébastien Bubeck, Zeyuan Allen-Zhu, 2018, ICML.

Physics of Language Models: Part 1, Context-Free Grammar

Yuanzhi Li, Zeyuan Allen-Zhu, 2023, ArXiv.

Algorithms and matching lower bounds for approximately-convex optimization

Yuanzhi Li, Andrej Risteski, Yuanzhi Li, 2016, NIPS.

A heuristic for statistical seriation

Nihar B. Shah, Yuanzhi Li, Komal Dhull, 2021, UAI.

Approximate maximum entropy principles via Goemans-Williamson with applications to provable variational methods

Yuanzhi Li, Andrej Risteski, Yuanzhi Li, 2016, NIPS.

Survey: No-Cloning Theorem

Yuanzhi Li, Yuanzhi Li, 2014 .

Sparsity, variance and curvature in multi-armed bandits pdf

Yuanzhi Li, Sébastien Bubeck, Michael B. Cohen, 2017, ALT.

Competitively chasing convex bodies

Yin Tat Lee, Yuanzhi Li, Sébastien Bubeck, 2018, STOC.

Towards Understanding Ensemble, Knowledge Distillation and Self-Distillation in Deep Learning

Yuanzhi Li, Zeyuan Allen-Zhu, 2020, International Conference on Learning Representations.

Can SGD Learn Recurrent Neural Networks with Provable Generalization?

Yuanzhi Li, Zeyuan Allen-Zhu, Yuanzhi Li, 2019, NeurIPS.

Provable Alternating Gradient Descent for Non-negative Matrix Factorization with Strong Correlations

Yuanzhi Li, Yingyu Liang, Yuanzhi Li, 2017, ICML.

Online Improper Learning with an Approximation Oracle

Wei Hu, Yuanzhi Li, Zhiyuan Li, 2018, NeurIPS.

Follow the Compressed Leader: Faster Algorithms for Matrix Multiplicative Weight Updates pdf

Yuanzhi Li, Zeyuan Allen Zhu, Yuanzhi Li, 2017, ArXiv.

Physics of Language Models: Part 1, Context-Free Grammar

Yuanzhi Li, Zeyuan Allen-Zhu, 2023, ArXiv.

The Well Tempered Lasso

Yuanzhi Li, Yoram Singer, Y. Singer, 2018, ICML.

Towards Understanding Clean Generalization and Robust Overfitting in Adversarial Training

Yuanzhi Li, Binghui Li, 2023, 2306.01271.

Toward Understanding the Feature Learning Process of Self-supervised Contrastive Learning pdf

Yuanzhi Li, Zixin Wen, Yuanzhi Li, 2021, ICML.

Online Improper Learning with an Approximation Oracle

Elad Hazan, Yuanzhi Li, Wei Hu, 2018, NeurIPS.

An homotopy method for $\ell_p$ regression provably beyond self-concordance and in input-sparsity time

Michael B. Cohen, Sébastien Bubeck, Yuanzhi Li, 2017, 1711.01328.

On the Convergence Rate of Training Recurrent Neural Networks

Yuanzhi Li, Zeyuan Allen-Zhu, Zhao Song, 2018, NeurIPS.

The Mechanism of Prediction Head in Non-contrastive Self-supervised Learning

Yuanzhi Li, Zixin Wen, 2022, NeurIPS.