AI Safety

Out-of-Distribution Detection With Reconstruction Error and Typicality-Based Penalty WACV 2023 Paraphrasing evades detectors of AI-generated text, but retrieval is an effective defense NeurIPS 2023 Part-Based Models Improve Adversarial Robustness ICLR 2023 PatchNAS: Repairing DNNs in Deployment with Patched Network Architecture Search AAAI 2023 Phase-aware Adversarial Defense for Improving Adversarial Robustness ICML 2023 Poisoning the Well: Can We Simultaneously Attack a Group of Learning Agents? IJCAI 2023 PolicyCleanse: Backdoor Detection and Mitigation for Competitive Reinforcement Learning ICCV 2023 Preprocessors Matter! Realistic Decision-Based Attacks on Machine Learning Systems ICML 2023 Preserving Privacy Through Dememorization: An Unlearning Technique For Mitigating Memorization Risks In Language Models EMNLP 2023 Privacy Aware Question-Answering System for Online Mental Health Risk Assessment ACL 2023 Privacy-Preserving Representations Are Not Enough: Recovering Scene Content From Camera Poses CVPR 2023 Proactive Deepfake Defence via Identity Watermarking WACV 2023 Probabilistic Categorical Adversarial Attack and Adversarial Training ICML 2023 Probabilistic Reasoning and Learning for Trustworthy AI AAAI 2023 Probabilistic Safeguard for Reinforcement Learning Using Safety Index Guided Gaussian Process Models L4DC 2023 Probabilistic Temporal Logic for Reasoning about Bounded Policies IJCAI 2023 Probabilities Are Not Enough: Formal Controller Synthesis for Stochastic Dynamical Models with Epistemic Uncertainty AAAI 2023 Progressive Backdoor Erasing via Connecting Backdoor and Adversarial Attacks CVPR 2023 Prompt as Triggers for Backdoor Attack: Examining the Vulnerability in Language Models EMNLP 2023 Provable Adversarial Robustness for Group Equivariant Tasks: Graphs, Point Clouds, Molecules, and More NeurIPS 2023 Provable Robustness against Wasserstein Distribution Shifts via Input Randomization ICLR 2023 Provable Safe Reinforcement Learning with Binary Feedback AISTATS 2023 Provably Bounding Neural Network Preimages NeurIPS 2023 Provably Efficient Primal-Dual Reinforcement Learning for CMDPs with Non-stationary Objectives and Constraints AAAI 2023 Provably Safe Reinforcement Learning with Step-wise Violation Constraints NeurIPS 2023