Soheil Feizi

papers

1,251

total citations

papers (25)

Benchmarking Deep Learning Interpretability in Time Series Predictions

NEURIPS 2020arXiv

212

citations

(De)Randomized Smoothing for Certifiable Defense against Patch Attacks

NEURIPS 2020arXiv

174

citations

Segment and Complete: Defending Object Detectors Against Adversarial Patch Attacks With Robust Patch Detection

CVPR 2022arXiv

119

citations

Robust Optimal Transport with Applications in Generative Modeling and Domain Adaptation

NEURIPS 2020arXiv

118

citations

Improving Deep Learning Interpretability by Saliency Guided Training

NEURIPS 2021arXiv

104

citations

Fast Adversarial Attacks on Language Models In One GPU Minute

ICML 2024arXiv

citations

A Comprehensive Study of Image Classification Model Sensitivity to Foregrounds, Backgrounds, and Visual Attributes

CVPR 2022arXiv

citations

Dual Manifold Adversarial Robustness: Defense against Lp and non-Lp Adversarial Attacks

NEURIPS 2020arXiv

citations

Sample Efficient Detection and Classification of Adversarial Attacks via Self-Supervised Embeddings

ICCV 2021arXiv

citations

Spuriosity Rankings: Sorting Data to Measure and Mitigate Biases

NEURIPS 2023arXiv

citations

Lethal Dose Conjecture on Data Poisoning

NEURIPS 2022arXiv

citations

Temporal Robustness against Data poisoning

NEURIPS 2023arXiv

citations

Improved techniques for deterministic l2 robustness

NEURIPS 2022arXiv

citations

Towards Improved Input Masking for Convolutional Neural Networks

ICCV 2023arXiv

citations

Diffused Redundancy in Pre-trained Representations

NEURIPS 2023arXiv

citations

Exploring Geometry of Blind Spots in Vision models

NEURIPS 2023arXiv

citations

Toward Efficient Robust Training against Union of $\ell_p$ Threat Models

NEURIPS 2022

citations

964 Measuring Self-Supervised Representation Quality for Downstream Classification Using Discriminative Features

AAAI 2024

citations

Hard ImageNet: Segmentations for Objects with Strong Spurious Cues

NEURIPS 2022

citations

Soheil Feizi

papers (25)

Benchmarking Deep Learning Interpretability in Time Series Predictions

(De)Randomized Smoothing for Certifiable Defense against Patch Attacks

Segment and Complete: Defending Object Detectors Against Adversarial Patch Attacks With Robust Patch Detection

Robust Optimal Transport with Applications in Generative Modeling and Domain Adaptation

Improving Deep Learning Interpretability by Saliency Guided Training

Fast Adversarial Attacks on Language Models In One GPU Minute

A Comprehensive Study of Image Classification Model Sensitivity to Foregrounds, Backgrounds, and Visual Attributes

Dual Manifold Adversarial Robustness: Defense against Lp and non-Lp Adversarial Attacks

Strong Baselines for Parameter-Efficient Few-Shot Fine-Tuning

Low Curvature Activations Reduce Overfitting in Adversarial Training

Certifying Confidence via Randomized Smoothing

CUDA: Convolution-Based Unlearnable Datasets

Explicit Tradeoffs between Adversarial and Natural Distributional Robustness

On Mechanistic Knowledge Localization in Text-to-Image Generative Models

Sample Efficient Detection and Classification of Adversarial Attacks via Self-Supervised Embeddings

Spuriosity Rankings: Sorting Data to Measure and Mitigate Biases

Lethal Dose Conjecture on Data Poisoning

Temporal Robustness against Data poisoning

Improved techniques for deterministic l2 robustness

Towards Improved Input Masking for Convolutional Neural Networks

Diffused Redundancy in Pre-trained Representations

Exploring Geometry of Blind Spots in Vision models

Toward Efficient Robust Training against Union of $\ell_p$ Threat Models

964 Measuring Self-Supervised Representation Quality for Downstream Classification Using Discriminative Features

Hard ImageNet: Segmentations for Objects with Strong Spurious Cues

papers (25)

Benchmarking Deep Learning Interpretability in Time Series Predictions

(De)Randomized Smoothing for Certifiable Defense against Patch Attacks

Segment and Complete: Defending Object Detectors Against Adversarial Patch Attacks With Robust Patch Detection

Robust Optimal Transport with Applications in Generative Modeling and Domain Adaptation

Improving Deep Learning Interpretability by Saliency Guided Training

Fast Adversarial Attacks on Language Models In One GPU Minute

A Comprehensive Study of Image Classification Model Sensitivity to Foregrounds, Backgrounds, and Visual Attributes

Dual Manifold Adversarial Robustness: Defense against Lp and non-Lp Adversarial Attacks

Strong Baselines for Parameter-Efficient Few-Shot Fine-Tuning

Low Curvature Activations Reduce Overfitting in Adversarial Training

Certifying Confidence via Randomized Smoothing

CUDA: Convolution-Based Unlearnable Datasets

Explicit Tradeoffs between Adversarial and Natural Distributional Robustness

On Mechanistic Knowledge Localization in Text-to-Image Generative Models

Sample Efficient Detection and Classification of Adversarial Attacks via Self-Supervised Embeddings

Spuriosity Rankings: Sorting Data to Measure and Mitigate Biases

Lethal Dose Conjecture on Data Poisoning

Temporal Robustness against Data poisoning

Improved techniques for deterministic l2 robustness

Towards Improved Input Masking for Convolutional Neural Networks

Diffused Redundancy in Pre-trained Representations

Exploring Geometry of Blind Spots in Vision models

Toward Efficient Robust Training against Union of $\ell_p$ Threat Models

964 Measuring Self-Supervised Representation Quality for Downstream Classification Using Discriminative Features

Hard ImageNet: Segmentations for Objects with Strong Spurious Cues