Picture for Gal Vardi

Gal Vardi

Benign Overfitting and Grokking in ReLU Networks for XOR Cluster Data

Oct 04, 2023
Viaarxiv icon

Noisy Interpolation Learning with Shallow Univariate ReLU Networks

Aug 01, 2023
Figure 1 for Noisy Interpolation Learning with Shallow Univariate ReLU Networks
Figure 2 for Noisy Interpolation Learning with Shallow Univariate ReLU Networks
Figure 3 for Noisy Interpolation Learning with Shallow Univariate ReLU Networks
Figure 4 for Noisy Interpolation Learning with Shallow Univariate ReLU Networks
Viaarxiv icon

Deconstructing Data Reconstruction: Multiclass, Weight Decay and General Losses

Jul 04, 2023
Figure 1 for Deconstructing Data Reconstruction: Multiclass, Weight Decay and General Losses
Figure 2 for Deconstructing Data Reconstruction: Multiclass, Weight Decay and General Losses
Figure 3 for Deconstructing Data Reconstruction: Multiclass, Weight Decay and General Losses
Figure 4 for Deconstructing Data Reconstruction: Multiclass, Weight Decay and General Losses
Viaarxiv icon

An Agnostic View on the Cost of Overfitting in Ridge Regression

Jun 22, 2023
Viaarxiv icon

Most Neural Networks Are Almost Learnable

May 30, 2023
Viaarxiv icon

Reconstructing Training Data from Multiclass Neural Networks

May 05, 2023
Figure 1 for Reconstructing Training Data from Multiclass Neural Networks
Figure 2 for Reconstructing Training Data from Multiclass Neural Networks
Figure 3 for Reconstructing Training Data from Multiclass Neural Networks
Figure 4 for Reconstructing Training Data from Multiclass Neural Networks
Viaarxiv icon

The Double-Edged Sword of Implicit Bias: Generalization vs. Robustness in ReLU Networks

Mar 02, 2023
Viaarxiv icon

Benign Overfitting in Linear Classifiers and Leaky ReLU Networks from KKT Conditions for Margin Maximization

Mar 02, 2023
Viaarxiv icon

Adversarial Examples Exist in Two-Layer ReLU Networks for Low Dimensional Data Manifolds

Mar 01, 2023
Figure 1 for Adversarial Examples Exist in Two-Layer ReLU Networks for Low Dimensional Data Manifolds
Figure 2 for Adversarial Examples Exist in Two-Layer ReLU Networks for Low Dimensional Data Manifolds
Figure 3 for Adversarial Examples Exist in Two-Layer ReLU Networks for Low Dimensional Data Manifolds
Figure 4 for Adversarial Examples Exist in Two-Layer ReLU Networks for Low Dimensional Data Manifolds
Viaarxiv icon

Efficiently Learning Neural Networks: What Assumptions May Suffice?

Feb 15, 2023
Viaarxiv icon