AI Safety

Can we trust the evaluation on ChatGPT? ACL 2023 CARE-MI: Chinese Benchmark for Misinformation Evaluation in Maternity and Infant Care NeurIPS 2023 CBD: A Certified Backdoor Detector Based on Local Dominant Probability NeurIPS 2023 CDTA: A Cross-Domain Transfer-Based Attack with Contrastive Learning AAAI 2023 CellDAM: User-Space, Rootless Detection and Mitigation for 5G Data Plane NSDI 2023 Certification of Distributional Individual Fairness NeurIPS 2023 Certified Defense for Content Based Image Retrieval WACV 2023 Certified Policy Smoothing for Cooperative Multi-Agent Reinforcement Learning AAAI 2023 Certified Robustness via Dynamic Margin Maximization and Improved Lipschitz Regularization NeurIPS 2023 Certifying Ensembles: A General Certification Theory with S-Lipschitzness ICML 2023 CGBA: Curvature-aware Geometric Black-box Attack ICCV 2023 Chameleon: Adapting to Peer Images for Planting Durable Backdoors in Federated Learning ICML 2023 ClarifyDelphi: Reinforced Clarification Questions with Defeasibility Rewards for Social and Moral Situations ACL 2023 CleanCLIP: Mitigating Data Poisoning Attacks in Multimodal Contrastive Learning ICCV 2023 Collaborative Alignment of NLP Models NeurIPS 2023 Computer says “No”: The Case Against Empathetic Conversational AI ACL 2023 Connecting Certified and Adversarial Training NeurIPS 2023 Constrained Decision Transformer for Offline Safe Reinforcement Learning ICML 2023 Constraint-Conditioned Policy Optimization for Versatile Safe Reinforcement Learning NeurIPS 2023 Contextual Conservative Interleaving Bandits ICML 2023 Contrastive Self-Supervised Learning Leads to Higher Adversarial Susceptibility AAAI 2023 Controlling the Extraction of Memorized Data from Large Language Models via Prompt-Tuning ACL 2023 Copyright Violations and Large Language Models EMNLP 2023 Correct-by-Construction Reinforcement Learning of Cardiac Pacemakers from Duration Calculus Requirements AAAI 2023 Corrupting Neuron Explanations of Deep Visual Features ICCV 2023