Safety

Constrained Update Projection Approach to Safe Policy Optimization NeurIPS 2022 Counterfactual harm NeurIPS 2022 DeformRS: Certifying Input Deformations with Randomized Smoothing AAAI 2022 Distillation of RL Policies with Formal Guarantees via Variational Abstraction of Markov Decision Processes AAAI 2022 Enhancing Safe Exploration Using Safety State Augmentation NeurIPS 2022 Exploring Safer Behaviors for Deep Reinforcement Learning AAAI 2022 Handling and Presenting Harmful Text in NLP Research EMNLP 2022 Increasing Confidence in Adversarial Robustness Evaluations NeurIPS 2022 Leashing the Inner Demons: Self-Detoxification for Language Models AAAI 2022 On the Safety of Conversational Models: Taxonomy, Dataset, and Benchmark ACL 2022 On the Safety of Interpretable Machine Learning: A Maximum Deviation Approach NeurIPS 2022 Penalized Proximal Policy Optimization for Safe Reinforcement Learning IJCAI 2022 Planning to Avoid Side Effects AAAI 2022 Preemptive Image Robustification for Protecting Users against Man-in-the-Middle Adversarial Attacks AAAI 2022 ProsocialDialog: A Prosocial Backbone for Conversational Agents EMNLP 2022 Provable Defense against Backdoor Policies in Reinforcement Learning NeurIPS 2022 Red Teaming Language Models with Language Models EMNLP 2022 Risk-Driven Design of Perception Systems NeurIPS 2022 Safe Online Convex Optimization with Unknown Linear Safety Constraints AAAI 2022 SaFeRDialogues: Taking Feedback Gracefully after Conversational Safety Failures ACL 2022 SafeText: A Benchmark for Exploring Physical Safety in Language Models EMNLP 2022 Shield Decentralization for Safe Multi-Agent Reinforcement Learning NeurIPS 2022 Stability Verification in Stochastic Control Systems via Neural Network Supermartingales AAAI 2022 Tight Neural Network Verification via Semidefinite Relaxations and Linear Reformulations AAAI 2022 Toward Robust Spiking Neural Network Against Adversarial Perturbation NeurIPS 2022