Picture for Rong Ge

Rong Ge

Clemson University

A Uniform Confidence Phenomenon in Deep Learning and its Implications for Calibration

Add code
Jun 01, 2023
Viaarxiv icon

Smoothing the Landscape Boosts the Signal for SGD: Optimal Sample Complexity for Learning Single Index Models

Add code
May 18, 2023
Viaarxiv icon

Depth Separation with Multilayer Mean-Field Networks

Add code
Apr 03, 2023
Figure 1 for Depth Separation with Multilayer Mean-Field Networks
Figure 2 for Depth Separation with Multilayer Mean-Field Networks
Viaarxiv icon

Do Transformers Parse while Predicting the Masked Word?

Add code
Mar 14, 2023
Figure 1 for Do Transformers Parse while Predicting the Masked Word?
Figure 2 for Do Transformers Parse while Predicting the Masked Word?
Figure 3 for Do Transformers Parse while Predicting the Masked Word?
Figure 4 for Do Transformers Parse while Predicting the Masked Word?
Viaarxiv icon

Hiding Data Helps: On the Benefits of Masking for Sparse Coding

Add code
Feb 24, 2023
Viaarxiv icon

Implicit Regularization Leads to Benign Overfitting for Sparse Linear Regression

Add code
Feb 01, 2023
Viaarxiv icon

Provably Learning Diverse Features in Multi-View Data with Midpoint Mixup

Add code
Oct 24, 2022
Viaarxiv icon

Understanding Edge-of-Stability Training Dynamics with a Minimalist Example

Add code
Oct 07, 2022
Figure 1 for Understanding Edge-of-Stability Training Dynamics with a Minimalist Example
Figure 2 for Understanding Edge-of-Stability Training Dynamics with a Minimalist Example
Figure 3 for Understanding Edge-of-Stability Training Dynamics with a Minimalist Example
Figure 4 for Understanding Edge-of-Stability Training Dynamics with a Minimalist Example
Viaarxiv icon

Plateau in Monotonic Linear Interpolation -- A "Biased" View of Loss Landscape for Deep Networks

Add code
Oct 03, 2022
Figure 1 for Plateau in Monotonic Linear Interpolation -- A "Biased" View of Loss Landscape for Deep Networks
Figure 2 for Plateau in Monotonic Linear Interpolation -- A "Biased" View of Loss Landscape for Deep Networks
Figure 3 for Plateau in Monotonic Linear Interpolation -- A "Biased" View of Loss Landscape for Deep Networks
Figure 4 for Plateau in Monotonic Linear Interpolation -- A "Biased" View of Loss Landscape for Deep Networks
Viaarxiv icon

A Regression Approach to Learning-Augmented Online Algorithms

Add code
May 25, 2022
Figure 1 for A Regression Approach to Learning-Augmented Online Algorithms
Viaarxiv icon