AI Safety

Decision-based Black-box Attack Against Vision Transformers via Patch-wise Adversarial Removal NeurIPS 2022 Defending against Model Stealing via Verifying Embedded External Features AAAI 2022 Defining and Characterizing Reward Gaming NeurIPS 2022 Delivering Trustworthy AI through Formal XAI AAAI 2022 Demystifying the Adversarial Robustness of Random Transformation Defenses ICML 2022 Detecting textual adversarial examples through randomized substitution and vote UAI 2022 Differential Assessment of Black-Box AI Agents AAAI 2022 Diffusion Models for Adversarial Purification ICML 2022 Dim-Krum: Backdoor-Resistant Federated Learning for NLP with Dimension-wise Krum-Based Aggregation EMNLP 2022 Direct Behavior Specification via Constrained Reinforcement Learning ICML 2022 Diving Deep into Modes of Fact Hallucinations in Dialogue Systems EMNLP 2022 Double Bubble, Toil and Trouble: Enhancing Certified Robustness through Transitivity NeurIPS 2022 Double Sampling Randomized Smoothing ICML 2022 Enhancing Adversarial Robustness for Deep Metric Learning CVPR 2022 Enhancing Safe Exploration Using Safety State Augmentation NeurIPS 2022 Error Amplification When Updating Deployed Machine Learning Models MLHC 2022 EvenNet: Ignoring Odd-Hop Neighbors Improves Robustness of Graph Neural Networks NeurIPS 2022 Evolution of Neural Tangent Kernels under Benign and Adversarial Training NeurIPS 2022 Exploring Safer Behaviors for Deep Reinforcement Learning AAAI 2022 Exploring the Universal Vulnerability of Prompt-based Learning Paradigm NAACL 2022 Exploring the Vulnerability of Deep Reinforcement Learning-based Emergency Control for Low Carbon Power Systems IJCAI 2022 Expose Backdoors on the Way: A Feature-Based Efficient Defense against Textual Backdoor Attacks EMNLP 2022 Fast and Reliable Evaluation of Adversarial Robustness with Minimum-Margin Attack ICML 2022 FedInv: Byzantine-Robust Federated Learning by Inversing Local Model Updates AAAI 2022 Fine-mixing: Mitigating Backdoors in Fine-tuned Language Models EMNLP 2022