Research map / Adversarial robustness and security / Certified and empirical robustness
Bayesian Learning with Information Gain Provably Bounds Risk for a Robust Adversarial Defense
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use IG-BNN instead of Adv-BNN | VGG-16 · CIFAR-10, ε=0.035 | accuracy: 37.7 -> 50.2 | Helped |
| Use IG-BNN instead of Adv-BNN | ModelA · STL-10, ε=0.035 | accuracy: 31.4 -> 48.2 | Helped |
| Use IG-BNN instead of Adv. Training | VGG-16 · CIFAR-10, ε=0.035 | accuracy: 31.1 -> 50.2 | Helped |
| Use IG-BNN instead of Adv. Training | ModelA · STL-10, ε=0.035 | accuracy: 27.4 -> 48.2 | Helped |
| Add SVGD instead of BBB inference (ablation) | Adversarially trained BNN on STL-10 · STL-10, ε=0.035 | accuracy: 31.4 -> 36.6 | Helped |
| Add the IG objective (ablation) | Adversarially trained SVGD BNN on STL-10 · STL-10, ε=0.035 | accuracy: 36.6 -> 48.2 | Helped |
| Maximize rather than minimize the IG objective (ablation) | BNN on STL-10 · STL-10 | accuracy | Hurt |
Closest papers by meaning
- Adv-BNN: Improved Adversarial Defense through Robust Bayesian Neural Network (ICLR 2019 · similarity 0.93)
- Defense Through Diverse Directions (ICML 2020 · similarity 0.92)
- Robustness of Bayesian Neural Networks to Gradient-Based Attacks (NeurIPS 2020 · similarity 0.90)
- Bayesian Adversarial Learning (NeurIPS 2018 · similarity 0.90)
- Adversarial Training Can Provably Improve Robustness: Theoretical Analysis of Feature Learning Process Under Structured Data (ICLR 2025 · similarity 0.89)
- Adversarial Robustness with Semi-Infinite Constrained Learning (NeurIPS 2021 · similarity 0.89)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.