Search Results for author:
Found 8 papers, 7 papers with code
Placing Objects in Context via Inpainting for Out-of-distribution Segmentation
In our experiments, we present different anomaly segmentation datasets based on POC-generated data and show that POC can improve the performance of recent state-of-the-art anomaly fine-tuning methods in several standardized benchmarks.
A Survey on Transferability of Adversarial Examples across Deep Neural Networks
This survey explores the landscape of the adversarial transferability of adversarial examples.
Reliability in Semantic Segmentation: Are We on the Right Track?
We analyze a broad variety of models, spanning from older ResNet-based architectures to novel transformers and assess their reliability based on four metrics: robustness, calibration, misclassification detection and out-of-distribution (OOD) detection.
Catastrophic overfitting can be induced with discriminative non-robust features
Through extensive experiments we analyze this novel phenomenon and discover that the presence of these easy features induces a learning shortcut that leads to CO. Our findings provide new insights into the mechanisms of CO and improve our understanding of the dynamics of AT.
On the Road to Online Adaptation for Semantic Image Segmentation
We propose a new problem formulation and a corresponding evaluation framework to advance research on unsupervised domain adaptation for semantic image segmentation.
Make Some Noise: Reliable and Efficient Single-Step Adversarial Training
Recently, Wong et al. showed that adversarial training with single-step FGSM leads to a characteristic failure mode named Catastrophic Overfitting (CO), in which a model becomes suddenly vulnerable to multi-step attacks.
Towards fast and effective single-step adversarial training
In this work, we methodically revisit the role of noise and clipping in single-step adversarial training.
Progressive Skeletonization: Trimming more fat from a network at initialization
Recent studies have shown that skeletonization (pruning parameters) of networks \textit{at initialization} provides all the practical benefits of sparsity both at inference and training time, while only marginally degrading their performance.
