AI Safety

Deep Verifier Networks: Verification of Deep Discriminative Models with Deep Generative Models AAAI 2021 Defending Against Universal Adversarial Patches by Clipping Feature Norms ICCV 2021 DeHiB: Deep Hidden Backdoor Attack on Semi-supervised Learning via Adversarial Perturbation AAAI 2021 Density Constrained Reinforcement Learning ICML 2021 Detection and Defense of Topological Adversarial Attacks on Graphs AISTATS 2021 Disrupting Deep Uncertainty Estimation Without Harming Accuracy NeurIPS 2021 Drawing Robust Scratch Tickets: Subnetworks with Inborn Robustness Are Found within Randomly Initialized Networks NeurIPS 2021 EarFisher: Detecting Wireless Eavesdroppers by Stimulating and Sensing Memory EMR NSDI 2021 Efficient Certified Defenses Against Patch Attacks on Image Classifiers ICLR 2021 Enforcing robust control guarantees within neural network policies ICLR 2021 Ethically Compliant Sequential Decision Making AAAI 2021 Evaluating Gradient Inversion Attacks and Defenses in Federated Learning NeurIPS 2021 Excess Capacity and Backdoor Poisoning NeurIPS 2021 Exploiting Sparsity for Neural Network Verification L4DC 2021 Exploring Architectural Ingredients of Adversarially Robust Deep Neural Networks NeurIPS 2021 Exploring the Vulnerability of Deep Neural Networks: A Study of Parameter Corruption AAAI 2021 Fair and Truthful Mechanisms for Dichotomous Valuations AAAI 2021 Fast Geometric Projections for Local Robustness Certification ICLR 2021 Fast Training of Provably Robust Neural Networks by SingleProp AAAI 2021 Fault-Tolerant Federated Reinforcement Learning with Theoretical Guarantee NeurIPS 2021 Fooling a Complete Neural Network Verifier ICLR 2021 Fooling Thermal Infrared Pedestrian Detectors in Real World Using Small Bulbs AAAI 2021 Formal verification of neural networks for safety-critical tasks in deep reinforcement learning UAI 2021 Foundations of Symbolic Languages for Model Interpretability NeurIPS 2021 Gaussian Process-Based Real-Time Learning for Safety Critical Applications ICML 2021