Multimodal Safety-Critical Scenarios Generation for Decision-Making Algorithms Evaluation
arXiv:2009.08311
Abstract
Existing neural network-based autonomous systems are shown to be vulnerable against adversarial attacks, therefore sophisticated evaluation on their robustness is of great importance. However, evaluating the robustness only under the worst-case scenarios based on known attacks is not comprehensive, not to mention that some of them even rarely occur in the real world. In addition, the distribution of safety-critical data is usually multimodal, while most traditional attacks and evaluation methods focus on a single modality. To solve the above challenges, we propose a flow-based multimodal safety-critical scenario generator for evaluating decisionmaking algorithms. The proposed generative model is optimized with weighted likelihood maximization and a gradient-based sampling procedure is integrated to improve the sampling efficiency. The safety-critical scenarios are generated by querying the task algorithms and the log-likelihood of the generated scenarios is in proportion to the risk level. Experiments on a self-driving task demonstrate our advantages in terms of testing efficiency and multimodal modeling capability. We evaluate six Reinforcement Learning algorithms with our generated traffic scenarios and provide empirical conclusions about their robustness.
arXiv comment8 pages, 7 figures
References in corpus (21)
- Continuous control with deep reinforcement learning
- Playing Atari with Deep Reinforcement Learning
- Soft Actor-Critic: Off-Policy Maximum Entropy Deep Reinforcement Learning with a Stochastic Actor
- ZOO: Zeroth Order Optimization based Black-box Attacks to Deep Neural Networks without Training Substitute Models
- Density estimation using Real NVP
- VIME: Variational Information Maximizing Exploration
- Model-Based Reinforcement Learning for Atari
- Black-box Adversarial Attacks with Limited Queries and Information
- CARLA: An Open Urban Driving Simulator
- Benchmarking Batch Deep Reinforcement Learning Algorithms
- Model-Based Active Exploration
- Learning Likelihoods with Conditional Normalizing Flows
- Efficient Black-box Assessment of Autonomous Vehicle Safety
- Self-Supervised Exploration via Disagreement
- A New Multi-vehicle Trajectory Generator to Simulate Vehicle-to-Vehicle Encounters
- Potential adversarial samples for white-box attacks
- Adversarial Evaluation of Autonomous Vehicles in Lane-Change Scenarios
- Learning To Simulate
- Learning to Collide: An Adaptive Safety-Critical Scenarios Generating Method
- CMTS: Conditional Multiple Trajectory Synthesizer for Generating Safety-critical Driving Scenarios
- Implicit Generative Modeling for Efficient Exploration