Picture for Michael Spratling

Michael Spratling

One Prompt Word is Enough to Boost Adversarial Robustness for Pre-trained Vision-Language Models

Add code
Mar 04, 2024
Figure 1 for One Prompt Word is Enough to Boost Adversarial Robustness for Pre-trained Vision-Language Models
Figure 2 for One Prompt Word is Enough to Boost Adversarial Robustness for Pre-trained Vision-Language Models
Figure 3 for One Prompt Word is Enough to Boost Adversarial Robustness for Pre-trained Vision-Language Models
Figure 4 for One Prompt Word is Enough to Boost Adversarial Robustness for Pre-trained Vision-Language Models
Viaarxiv icon

The Importance of Anti-Aliasing in Tiny Object Detection

Add code
Oct 22, 2023
Figure 1 for The Importance of Anti-Aliasing in Tiny Object Detection
Figure 2 for The Importance of Anti-Aliasing in Tiny Object Detection
Figure 3 for The Importance of Anti-Aliasing in Tiny Object Detection
Figure 4 for The Importance of Anti-Aliasing in Tiny Object Detection
Viaarxiv icon

OODRobustBench: benchmarking and analyzing adversarial robustness under distribution shift

Add code
Oct 19, 2023
Figure 1 for OODRobustBench: benchmarking and analyzing adversarial robustness under distribution shift
Figure 2 for OODRobustBench: benchmarking and analyzing adversarial robustness under distribution shift
Figure 3 for OODRobustBench: benchmarking and analyzing adversarial robustness under distribution shift
Figure 4 for OODRobustBench: benchmarking and analyzing adversarial robustness under distribution shift
Viaarxiv icon

When Multi-Task Learning Meets Partial Supervision: A Computer Vision Review

Jul 25, 2023
Figure 1 for When Multi-Task Learning Meets Partial Supervision: A Computer Vision Review
Figure 2 for When Multi-Task Learning Meets Partial Supervision: A Computer Vision Review
Figure 3 for When Multi-Task Learning Meets Partial Supervision: A Computer Vision Review
Figure 4 for When Multi-Task Learning Meets Partial Supervision: A Computer Vision Review
Viaarxiv icon

AROID: Improving Adversarial Robustness through Online Instance-wise Data Augmentation

Jun 12, 2023
Figure 1 for AROID: Improving Adversarial Robustness through Online Instance-wise Data Augmentation
Figure 2 for AROID: Improving Adversarial Robustness through Online Instance-wise Data Augmentation
Figure 3 for AROID: Improving Adversarial Robustness through Online Instance-wise Data Augmentation
Figure 4 for AROID: Improving Adversarial Robustness through Online Instance-wise Data Augmentation
Viaarxiv icon

Improved Adversarial Training Through Adaptive Instance-wise Loss Smoothing

Add code
Mar 27, 2023
Figure 1 for Improved Adversarial Training Through Adaptive Instance-wise Loss Smoothing
Figure 2 for Improved Adversarial Training Through Adaptive Instance-wise Loss Smoothing
Figure 3 for Improved Adversarial Training Through Adaptive Instance-wise Loss Smoothing
Figure 4 for Improved Adversarial Training Through Adaptive Instance-wise Loss Smoothing
Viaarxiv icon

Rethinking the backbone architecture for tiny object detection

Mar 20, 2023
Figure 1 for Rethinking the backbone architecture for tiny object detection
Figure 2 for Rethinking the backbone architecture for tiny object detection
Figure 3 for Rethinking the backbone architecture for tiny object detection
Figure 4 for Rethinking the backbone architecture for tiny object detection
Viaarxiv icon

Data Augmentation Alone Can Improve Adversarial Training

Add code
Jan 24, 2023
Figure 1 for Data Augmentation Alone Can Improve Adversarial Training
Figure 2 for Data Augmentation Alone Can Improve Adversarial Training
Figure 3 for Data Augmentation Alone Can Improve Adversarial Training
Figure 4 for Data Augmentation Alone Can Improve Adversarial Training
Viaarxiv icon

Understanding and Combating Robust Overfitting via Input Loss Landscape Analysis and Regularization

Add code
Dec 09, 2022
Figure 1 for Understanding and Combating Robust Overfitting via Input Loss Landscape Analysis and Regularization
Figure 2 for Understanding and Combating Robust Overfitting via Input Loss Landscape Analysis and Regularization
Figure 3 for Understanding and Combating Robust Overfitting via Input Loss Landscape Analysis and Regularization
Figure 4 for Understanding and Combating Robust Overfitting via Input Loss Landscape Analysis and Regularization
Viaarxiv icon

On the biological plausibility of orthogonal initialisation for solving gradient instability in deep neural networks

Add code
Oct 27, 2022
Figure 1 for On the biological plausibility of orthogonal initialisation for solving gradient instability in deep neural networks
Figure 2 for On the biological plausibility of orthogonal initialisation for solving gradient instability in deep neural networks
Figure 3 for On the biological plausibility of orthogonal initialisation for solving gradient instability in deep neural networks
Figure 4 for On the biological plausibility of orthogonal initialisation for solving gradient instability in deep neural networks
Viaarxiv icon