AI Safety

Automatically Exposing Problems with Neural Dialog Models EMNLP 2021 Backdoor Attacks Against Deep Learning Systems in the Physical World CVPR 2021 Backdoor Attacks on Pre-trained Models by Layerwise Weight Poisoning EMNLP 2021 Backdoor Attack with Imperceptible Input and Latent Modification NeurIPS 2021 Backdoor Scanning for Deep Neural Networks through K-Arm Optimization ICML 2021 Bayesian Inference with Certifiable Adversarial Robustness AISTATS 2021 Be Careful about Poisoned Word Embeddings: Exploring the Vulnerability of the Embedding Layers in NLP Models NAACL 2021 BFClass: A Backdoor-free Text Classification Framework EMNLP 2021 CARTL: Cooperative Adversarially-Robust Transfer Learning ICML 2021 Center Smoothing: Certified Robustness for Networks with Structured Outputs NeurIPS 2021 Certification of iterative predictions in Bayesian neural networks UAI 2021 Certified Robustness to Programmable Transformations in LSTMs EMNLP 2021 Certified Robustness to Word Substitution Attack with Differential Privacy NAACL 2021 Certifying Incremental Quadratic Constraints for Neural Networks via Convex Optimization L4DC 2021 Certify or Predict: Boosting Certified Robustness with Compositional Architectures ICLR 2021 CLEAR: Clean-Up Sample-Targeted Backdoor in Neural Networks ICCV 2021 Closing the BIG-LID: An Effective Local Intrinsic Dimensionality Defense for Nonlinear Regression Poisoning IJCAI 2021 Collective Robustness Certificates: Exploiting Interdependence in Graph Neural Networks ICLR 2021 Common Sense Bias in Semantic Role Labeling EMNLP 2021 Concealed Data Poisoning Attacks on NLP Models NAACL 2021 Counterexample Guided RL Policy Refinement Using Bayesian Optimization NeurIPS 2021 CRPO: A New Approach for Safe Reinforcement Learning with Convergence Guarantee ICML 2021 Data Efficient Algorithms and Interpretability Requirements for Personalized Assessment of Taskable AI Systems IJCAI 2021 Deep Feature Space Trojan Attack of Neural Networks by Controlled Detoxification AAAI 2021 DEEPSPLIT: An Efficient Splitting Method for Neural Network Verification via Indirect Effect Analysis IJCAI 2021