robust-ood-detection
Robust Out-of-distribution Detection in Neural Networks
- 73
- Python
- Apache-2.0
- 5 y ago
jfc43
Research Scientist at Google.
Built from jfc43's public GitHub profile, Sunnyvale. Not on Take My Repo yet, so there are no posts or updates here.
Robust Out-of-distribution Detection in Neural Networks
We propose a theoretically motivated method, Adversarial Training with informative Outlier Mining (ATOM), which improves the robustness of OOD detection to various types of adversarial OOD inputs and establishes state-of-the-art performance.
Propose a principled and practically effective framework for unsupervised accuracy estimation and error detection tasks with theoretical analysis and state-of-the-art performance.
Robust Attribution Regularization
Pixel discretization defense against adversarial attacks
Exploring evaluating the adversarial robustness of transductive-learning based defenses.
Study adversarially-robust classification with rejection where rejection has cost monotonically non-increasing in the perturbation magnitude, and propose a novel defense method with strong performance.