ICML 2023 accepted papers on Adversarial robustness and security
56 of the 1,865 papers accepted at ICML 2023 (3.0%) are on Adversarial robustness and security, compared with 4.6% at ICML 2022. They fall into 4 topics.
See ICML 2023 in the interactive map
Working on something in this area? Describe your idea in scime atlas to see the closest accepted papers by meaning and how crowded the spot has become.
Topics and papers
Certified and empirical robustness · 32 papers
- Better Diffusion Models Further Improve Adversarial Training (52 citations)
- Run-off Election: Improved Provable Defense against Data Poisoning Attacks (5 citations)
- Interpolation for Robust Learning: Data Augmentation on Wasserstein Geodesics (3 citations)
- On the Robustness of Randomized Ensembles to Adversarial Perturbations (3 citations)
- Beyond the Universal Law of Robustness: Sharper Laws for Random Features and Neural Tangent Kernels (2 citations)
- Certified Robust Neural Networks: Generalization and Corruption Resistance (2 citations)
- NeRFool: Uncovering the Vulnerability of Generalizable Neural Radiance Fields against Adversarial Perturbations (2 citations)
- One-vs-the-Rest Loss to Focus on Important Samples in Adversarial Training (2 citations)
- Robust One-Class Classification with Signed Distance Function using 1-Lipschitz Neural Networks (2 citations)
- Understanding the Impact of Adversarial Robustness on Accuracy Disparity (2 citations)
- Adversarial Parameter Attack on Deep Neural Networks (1 citation)
- GAT: Guided Adversarial Training with Pareto-optimal Auxiliary Tasks (1 citation)
and 20 more in the interactive map.
Attacks on text and graphs · 18 papers
- Adversarial Example Does Good: Preventing Painting Imitation from Diffusion Models via Adversarial Examples (19 citations)
- Adversarial Policies Beat Superhuman Go AIs (9 citations)
- Data Poisoning Attacks Against Multimodal Encoders (7 citations)
- Detecting Adversarial Data by Probing Multiple Perturbations Using Expected Perturbation Score (4 citations)
- Adversarial Cheap Talk (2 citations)
- Robust Perception through Equivariance (2 citations)
- NNSplitter: An Active Defense Solution for DNN Model via Automated Weight Obfuscation (1 citation)
- Exploring Model Dynamics for Accumulative Poisoning Discovery
- Exploring the Limits of Model-Targeted Indiscriminate Data Poisoning Attacks
- Probabilistic Categorical Adversarial Attack and Adversarial Training
- Uncovering Adversarial Risks of Test-Time Adaptation
- Deep Perturbation Learning: Enhancing the Network Performance via Image Perturbations
and 6 more in the interactive map.