Get our free extension to see links to code for papers anywhere online!

Chrome logo Add to Chrome

Firefox logo Add to Firefox

Picture for Sanjeev Arora

Rip van Winkle's Razor: A Simple Estimate of Overfit to Test Data


Feb 25, 2021
Sanjeev Arora, Yi Zhang


  Access Paper or Ask Questions

On the Validity of Modeling SGD with Stochastic Differential Equations (SDEs)


Feb 24, 2021
Zhiyuan Li, Sadhika Malladi, Sanjeev Arora

* 30 pages, 19 figures 

  Access Paper or Ask Questions

Why Are Convolutional Nets More Sample-Efficient than Fully-Connected Nets?


Oct 16, 2020
Zhiyuan Li, Yi Zhang, Sanjeev Arora

* 24 pages, 1 figure 

  Access Paper or Ask Questions

TextHide: Tackling Data Privacy in Language Understanding Tasks


Oct 12, 2020
Yangsibo Huang, Zhao Song, Danqi Chen, Kai Li, Sanjeev Arora

* Findings of EMNLP 2020 

  Access Paper or Ask Questions

A Mathematical Exploration of Why Language Models Help Solve Downstream Tasks


Oct 07, 2020
Nikunj Saunshi, Sadhika Malladi, Sanjeev Arora

* 29 pages 

  Access Paper or Ask Questions

Reconciling Modern Deep Learning with Traditional Optimization Analyses: The Intrinsic Learning Rate


Oct 06, 2020
Zhiyuan Li, Kaifeng Lyu, Sanjeev Arora

* 25 pages, 12 figures. Accepted By 34th Conference on Neural Information Processing Systems (NeurIPS 2020) 

  Access Paper or Ask Questions

InstaHide: Instance-hiding Schemes for Private Distributed Learning


Oct 06, 2020
Yangsibo Huang, Zhao Song, Kai Li, Sanjeev Arora

* ICML 2020 

  Access Paper or Ask Questions

Privacy-preserving Learning via Deep Net Pruning


Mar 04, 2020
Yangsibo Huang, Yushan Su, Sachin Ravi, Zhao Song, Sanjeev Arora, Kai Li


  Access Paper or Ask Questions

A Sample Complexity Separation between Non-Convex and Convex Meta-Learning


Feb 25, 2020
Nikunj Saunshi, Yi Zhang, Mikhail Khodak, Sanjeev Arora

* 34 pages 

  Access Paper or Ask Questions

Provable Representation Learning for Imitation Learning via Bi-level Optimization


Feb 24, 2020
Sanjeev Arora, Simon S. Du, Sham Kakade, Yuping Luo, Nikunj Saunshi

* 26 pages 

  Access Paper or Ask Questions

Over-parameterized Adversarial Training: An Analysis Overcoming the Curse of Dimensionality


Feb 24, 2020
Yi Zhang, Orestis Plevrakis, Simon S. Du, Xingguo Li, Zhao Song, Sanjeev Arora


  Access Paper or Ask Questions

An Exponential Learning Rate Schedule for Deep Learning


Nov 21, 2019
Zhiyuan Li, Sanjeev Arora


  Access Paper or Ask Questions

Enhanced Convolutional Neural Tangent Kernels


Nov 03, 2019
Zhiyuan Li, Ruosong Wang, Dingli Yu, Simon S. Du, Wei Hu, Ruslan Salakhutdinov, Sanjeev Arora


  Access Paper or Ask Questions

Harnessing the Power of Infinitely Wide Deep Nets on Small-data Tasks


Oct 27, 2019
Sanjeev Arora, Simon S. Du, Zhiyuan Li, Ruslan Salakhutdinov, Ruosong Wang, Dingli Yu

* Code for UCI experiments: https://github.com/LeoYu/neural-tangent-kernel-UCI 

  Access Paper or Ask Questions

Explaining Landscape Connectivity of Low-cost Solutions for Multilayer Nets


Jun 14, 2019
Rohith Kuditipudi, Xiang Wang, Holden Lee, Yi Zhang, Zhiyuan Li, Wei Hu, Sanjeev Arora, Rong Ge


  Access Paper or Ask Questions

A Simple Saliency Method That Passes the Sanity Checks


Jun 07, 2019
Arushi Gupta, Sanjeev Arora

* Small typo on paragraph 3 of section 3 fixed 

  Access Paper or Ask Questions

Implicit Regularization in Deep Matrix Factorization


Jun 04, 2019
Sanjeev Arora, Nadav Cohen, Wei Hu, Yuping Luo


  Access Paper or Ask Questions

On Exact Computation with an Infinitely Wide Neural Net


Apr 26, 2019
Sanjeev Arora, Simon S. Du, Wei Hu, Zhiyuan Li, Ruslan Salakhutdinov, Ruosong Wang

* abstract shortened to meet the constraint 

  Access Paper or Ask Questions

A Theoretical Analysis of Contrastive Unsupervised Representation Learning


Feb 25, 2019
Sanjeev Arora, Hrishikesh Khandeparkar, Mikhail Khodak, Orestis Plevrakis, Nikunj Saunshi

* 19 pages, 5 figures 

  Access Paper or Ask Questions

Fine-Grained Analysis of Optimization and Generalization for Overparameterized Two-Layer Neural Networks


Jan 24, 2019
Sanjeev Arora, Simon S. Du, Wei Hu, Zhiyuan Li, Ruosong Wang


  Access Paper or Ask Questions

Theoretical Analysis of Auto Rate-Tuning by Batch Normalization


Dec 10, 2018
Sanjeev Arora, Zhiyuan Li, Kaifeng Lyu

* 22 pages 

  Access Paper or Ask Questions

Stronger generalization bounds for deep nets via a compression approach


Nov 05, 2018
Sanjeev Arora, Rong Ge, Behnam Neyshabur, Yi Zhang


  Access Paper or Ask Questions

Linear Algebraic Structure of Word Senses, with Applications to Polysemy


Jul 20, 2018
Sanjeev Arora, Yuanzhi Li, Yingyu Liang, Tengyu Ma, Andrej Risteski

* Appear in the Transactions of the Association for Computational Linguistics 2018, link: https://transacl.org/ojs/index.php/tacl/article/view/1346 

  Access Paper or Ask Questions

On the Optimization of Deep Networks: Implicit Acceleration by Overparameterization


Jun 11, 2018
Sanjeev Arora, Nadav Cohen, Elad Hazan

* Published at the International Conference on Machine Learning (ICML) 2018 

  Access Paper or Ask Questions