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

Add to Chrome

Add to Firefox

Picture for Rong Ge

Rong Ge

Understanding Deflation Process in Over-parametrized Tensor Decomposition

Jun 11, 2021
Rong Ge, Yunwei Ren, Xiang Wang, Mo Zhou

Figure 1 for Understanding Deflation Process in Over-parametrized Tensor Decomposition

Figure 2 for Understanding Deflation Process in Over-parametrized Tensor Decomposition

Figure 3 for Understanding Deflation Process in Over-parametrized Tensor Decomposition

Figure 4 for Understanding Deflation Process in Over-parametrized Tensor Decomposition

Access Paper or Ask Questions

A Local Convergence Theory for Mildly Over-Parameterized Two-Layer Neural Network

Feb 04, 2021
Mo Zhou, Rong Ge, Chi Jin

Figure 1 for A Local Convergence Theory for Mildly Over-Parameterized Two-Layer Neural Network

Figure 2 for A Local Convergence Theory for Mildly Over-Parameterized Two-Layer Neural Network

Figure 3 for A Local Convergence Theory for Mildly Over-Parameterized Two-Layer Neural Network

Figure 4 for A Local Convergence Theory for Mildly Over-Parameterized Two-Layer Neural Network

Access Paper or Ask Questions

Beyond Lazy Training for Over-parameterized Tensor Decomposition

Oct 22, 2020
Xiang Wang, Chenwei Wu, Jason D. Lee, Tengyu Ma, Rong Ge

Figure 1 for Beyond Lazy Training for Over-parameterized Tensor Decomposition

* NeurIPS 2020; the first two authors contribute equally

Access Paper or Ask Questions

Dissecting Hessian: Understanding Common Structure of Hessian in Neural Networks

Oct 08, 2020
Yikai Wu, Xingyu Zhu, Chenwei Wu, Annie Wang, Rong Ge

Figure 1 for Dissecting Hessian: Understanding Common Structure of Hessian in Neural Networks

Figure 2 for Dissecting Hessian: Understanding Common Structure of Hessian in Neural Networks

Figure 3 for Dissecting Hessian: Understanding Common Structure of Hessian in Neural Networks

Figure 4 for Dissecting Hessian: Understanding Common Structure of Hessian in Neural Networks

* 29 pages, 26 figures. Main text: 8 pages, 6 figures. First two authors have equal contribution and are in alphabetical order

Access Paper or Ask Questions

Efficient sampling from the Bingham distribution

Sep 30, 2020
Rong Ge, Holden Lee, Jianfeng Lu, Andrej Risteski

Access Paper or Ask Questions

Guarantees for Tuning the Step Size using a Learning-to-Learn Approach

Jun 30, 2020
Xiang Wang, Shuai Yuan, Chenwei Wu, Rong Ge

Figure 1 for Guarantees for Tuning the Step Size using a Learning-to-Learn Approach

Figure 2 for Guarantees for Tuning the Step Size using a Learning-to-Learn Approach

Figure 3 for Guarantees for Tuning the Step Size using a Learning-to-Learn Approach

Figure 4 for Guarantees for Tuning the Step Size using a Learning-to-Learn Approach

Access Paper or Ask Questions

Optimization Landscape of Tucker Decomposition

Jun 29, 2020
Abraham Frandsen, Rong Ge

Figure 1 for Optimization Landscape of Tucker Decomposition

Access Paper or Ask Questions

Extracting Latent State Representations with Linear Dynamics from Rich Observations

Jun 29, 2020
Abraham Frandsen, Rong Ge

Figure 1 for Extracting Latent State Representations with Linear Dynamics from Rich Observations

Figure 2 for Extracting Latent State Representations with Linear Dynamics from Rich Observations

Figure 3 for Extracting Latent State Representations with Linear Dynamics from Rich Observations

Figure 4 for Extracting Latent State Representations with Linear Dynamics from Rich Observations

Access Paper or Ask Questions

Energy-Aware DNN Graph Optimization

May 12, 2020
Yu Wang, Rong Ge, Shuang Qiu

Figure 1 for Energy-Aware DNN Graph Optimization

Figure 2 for Energy-Aware DNN Graph Optimization

Figure 3 for Energy-Aware DNN Graph Optimization

Access Paper or Ask Questions

High-Dimensional Robust Mean Estimation via Gradient Descent

May 04, 2020
Yu Cheng, Ilias Diakonikolas, Rong Ge, Mahdi Soltanolkotabi

* Under submission to ICML'20

Access Paper or Ask Questions

Spectral Learning on Matrices and Tensors

Apr 16, 2020
Majid Janzamin, Rong Ge, Jean Kossaifi, Anima Anandkumar

Figure 1 for Spectral Learning on Matrices and Tensors

Figure 2 for Spectral Learning on Matrices and Tensors

Figure 3 for Spectral Learning on Matrices and Tensors

Figure 4 for Spectral Learning on Matrices and Tensors

* Foundations and Trends in Machine Learning: Vol. 12: No. 5-6, pp 393-536 (2019)

Access Paper or Ask Questions

Estimating Normalizing Constants for Log-Concave Distributions: Algorithms and Lower Bounds

Nov 08, 2019
Rong Ge, Holden Lee, Jianfeng Lu

* 45 pages

Access Paper or Ask Questions

Mildly Overparametrized Neural Nets can Memorize Training Data Efficiently

Sep 26, 2019
Rong Ge, Runzhe Wang, Haoyu Zhao

Figure 1 for Mildly Overparametrized Neural Nets can Memorize Training Data Efficiently

Figure 2 for Mildly Overparametrized Neural Nets can Memorize Training Data Efficiently

Figure 3 for Mildly Overparametrized Neural Nets can Memorize Training Data Efficiently

Figure 4 for Mildly Overparametrized Neural Nets can Memorize Training Data Efficiently

* 38 pages

Access Paper or Ask Questions

Faster Algorithms for High-Dimensional Robust Covariance Estimation

Jun 11, 2019
Yu Cheng, Ilias Diakonikolas, Rong Ge, David Woodruff

Access Paper or Ask Questions

Stabilized SVRG: Simple Variance Reduction for Nonconvex Optimization

May 01, 2019
Rong Ge, Zhize Li, Weiyao Wang, Xiang Wang

Figure 1 for Stabilized SVRG: Simple Variance Reduction for Nonconvex Optimization

Figure 2 for Stabilized SVRG: Simple Variance Reduction for Nonconvex Optimization

Figure 3 for Stabilized SVRG: Simple Variance Reduction for Nonconvex Optimization

Figure 4 for Stabilized SVRG: Simple Variance Reduction for Nonconvex Optimization

Access Paper or Ask Questions

The Step Decay Schedule: A Near Optimal, Geometrically Decaying Learning Rate Procedure

Apr 29, 2019
Rong Ge, Sham M. Kakade, Rahul Kidambi, Praneeth Netrapalli

Figure 1 for The Step Decay Schedule: A Near Optimal, Geometrically Decaying Learning Rate Procedure

Figure 2 for The Step Decay Schedule: A Near Optimal, Geometrically Decaying Learning Rate Procedure

Figure 3 for The Step Decay Schedule: A Near Optimal, Geometrically Decaying Learning Rate Procedure

Figure 4 for The Step Decay Schedule: A Near Optimal, Geometrically Decaying Learning Rate Procedure

* 25 pages, 5 tables, 5 figures

Access Paper or Ask Questions

Stochastic Gradient Descent Escapes Saddle Points Efficiently

Feb 13, 2019
Chi Jin, Praneeth Netrapalli, Rong Ge, Sham M. Kakade, Michael I. Jordan

Figure 1 for Stochastic Gradient Descent Escapes Saddle Points Efficiently

Figure 2 for Stochastic Gradient Descent Escapes Saddle Points Efficiently

Figure 3 for Stochastic Gradient Descent Escapes Saddle Points Efficiently

Figure 4 for Stochastic Gradient Descent Escapes Saddle Points Efficiently

Access Paper or Ask Questions

A Short Note on Concentration Inequalities for Random Vectors with SubGaussian Norm

Feb 11, 2019
Chi Jin, Praneeth Netrapalli, Rong Ge, Sham M. Kakade, Michael I. Jordan

Access Paper or Ask Questions

Understanding Composition of Word Embeddings via Tensor Decomposition

Feb 02, 2019
Abraham Frandsen, Rong Ge

Figure 1 for Understanding Composition of Word Embeddings via Tensor Decomposition

Figure 2 for Understanding Composition of Word Embeddings via Tensor Decomposition

Figure 3 for Understanding Composition of Word Embeddings via Tensor Decomposition

Figure 4 for Understanding Composition of Word Embeddings via Tensor Decomposition

Access Paper or Ask Questions

Simulated Tempering Langevin Monte Carlo II: An Improved Proof using Soft Markov Chain Decomposition

Nov 29, 2018
Rong Ge, Holden Lee, Andrej Risteski

Figure 1 for Simulated Tempering Langevin Monte Carlo II: An Improved Proof using Soft Markov Chain Decomposition

* Advances in Neural Information Processing Systems 31 (2018)
* 55 pages. arXiv admin note: text overlap with arXiv:1710.02736

Access Paper or Ask Questions

High-Dimensional Robust Mean Estimation in Nearly-Linear Time

Nov 23, 2018
Yu Cheng, Ilias Diakonikolas, Rong Ge

Figure 1 for High-Dimensional Robust Mean Estimation in Nearly-Linear Time

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

Figure 1 for Stronger generalization bounds for deep nets via a compression approach

Figure 2 for Stronger generalization bounds for deep nets via a compression approach

Figure 3 for Stronger generalization bounds for deep nets via a compression approach

Figure 4 for Stronger generalization bounds for deep nets via a compression approach

Access Paper or Ask Questions

Global Convergence of Policy Gradient Methods for the Linear Quadratic Regulator

Oct 21, 2018
Maryam Fazel, Rong Ge, Sham M. Kakade, Mehran Mesbahi

Figure 1 for Global Convergence of Policy Gradient Methods for the Linear Quadratic Regulator

Access Paper or Ask Questions

On the Local Minima of the Empirical Risk

Oct 17, 2018
Chi Jin, Lydia T. Liu, Rong Ge, Michael I. Jordan

Figure 1 for On the Local Minima of the Empirical Risk

Figure 2 for On the Local Minima of the Empirical Risk

Figure 3 for On the Local Minima of the Empirical Risk

Figure 4 for On the Local Minima of the Empirical Risk

* To appear in NIPS 2018

Access Paper or Ask Questions

Learning Two-layer Neural Networks with Symmetric Inputs

Oct 16, 2018
Rong Ge, Rohith Kuditipudi, Zhize Li, Xiang Wang

Figure 1 for Learning Two-layer Neural Networks with Symmetric Inputs

Figure 2 for Learning Two-layer Neural Networks with Symmetric Inputs

Figure 3 for Learning Two-layer Neural Networks with Symmetric Inputs

Figure 4 for Learning Two-layer Neural Networks with Symmetric Inputs

Access Paper or Ask Questions

Non-Convex Matrix Completion Against a Semi-Random Adversary

Sep 07, 2018
Yu Cheng, Rong Ge

* added references and fixed typos

Access Paper or Ask Questions

Matrix Completion has No Spurious Local Minimum

Jul 22, 2018
Rong Ge, Jason D. Lee, Tengyu Ma

Figure 1 for Matrix Completion has No Spurious Local Minimum

* NIPS'16 best student paper. fixed Theorem 2.3 in preliminary section in the previous version. The results are not affected

Access Paper or Ask Questions

Beyond Log-concavity: Provable Guarantees for Sampling Multi-modal Distributions using Simulated Tempering Langevin Monte Carlo

Nov 06, 2017
Rong Ge, Holden Lee, Andrej Risteski

Figure 1 for Beyond Log-concavity: Provable Guarantees for Sampling Multi-modal Distributions using Simulated Tempering Langevin Monte Carlo

Figure 2 for Beyond Log-concavity: Provable Guarantees for Sampling Multi-modal Distributions using Simulated Tempering Langevin Monte Carlo

Figure 3 for Beyond Log-concavity: Provable Guarantees for Sampling Multi-modal Distributions using Simulated Tempering Langevin Monte Carlo

* 53 pages

Access Paper or Ask Questions

Learning One-hidden-layer Neural Networks with Landscape Design

Nov 03, 2017
Rong Ge, Jason D. Lee, Tengyu Ma

Figure 1 for Learning One-hidden-layer Neural Networks with Landscape Design

Figure 2 for Learning One-hidden-layer Neural Networks with Landscape Design

Figure 3 for Learning One-hidden-layer Neural Networks with Landscape Design

Access Paper or Ask Questions