AI Safety

BppAttack: Stealthy and Efficient Trojan Attacks Against Deep Neural Networks via Image Quantization and Contrastive Adversarial Learning CVPR 2022 Calibrated Learning to Defer with One-vs-All Classifiers ICML 2022 Can Adversarial Training Be Manipulated By Non-Robust Features? NeurIPS 2022 Can You Spot the Chameleon? Adversarially Camouflaging Images From Co-Salient Object Detection CVPR 2022 Capturing Failures of Large Language Models via Human Cognitive Biases NeurIPS 2022 CC-CERT: A Probabilistic Approach to Certify General Robustness of Neural Networks AAAI 2022 CerDEQ: Certifiable Deep Equilibrium Model ICML 2022 Certified Adversarial Robustness Under the Bounded Support Set ICML 2022 Certified Patch Robustness via Smoothed Vision Transformers CVPR 2022 Certified Robustness Against Natural Language Attacks by Causal Intervention ICML 2022 Certified Robustness via Locally Biased Randomized Smoothing L4DC 2022 Certified Robustness via Randomized Smoothing over Multiplicative Parameters of Input Transformations IJCAI 2022 Characteristics of Harmful Text: Towards Rigorous Benchmarking of Language Models NeurIPS 2022 Combating Adversaries with Anti-adversaries AAAI 2022 Consent as a Foundation for Responsible Autonomy AAAI 2022 Conservative and Adaptive Penalty for Model-Based Safe Reinforcement Learning AAAI 2022 Constrained Update Projection Approach to Safe Policy Optimization NeurIPS 2022 Constructing Highly Inductive Contexts for Dialogue Safety through Controllable Reverse Generation EMNLP 2022 Contextualizing Language Models for Norms Diverging from Social Majority EMNLP 2022 Cosine Model Watermarking against Ensemble Distillation AAAI 2022 Counterfactual harm NeurIPS 2022 Cross-document Misinformation Detection based on Event Graph Reasoning NAACL 2022 Data dependent randomized smoothing UAI 2022 Data-Driven Safety Verification of Stochastic Systems via Barrier Certificates: A Wait-and-Judge Approach L4DC 2022 Data poisoning attacks on off-policy policy evaluation methods UAI 2022