arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2009.08311cs.LGcs.ROstat.ML
收藏

用于决策算法评估的多模态安全关键场景生成

Multimodal Safety-Critical Scenarios Generation for Decision-Making Algorithms Evaluation

Carnegie Mellon University(卡内基梅隆大学);UIUC(伊利诺伊大学厄巴纳-香槟分校);Robert Bosch LLC(罗伯特·博世有限责任公司)

Wenhao Ding, Baiming Chen, Bo Li, Kim Ji Eun, Ding Zhao

摘要:

现有的基于神经网络的自治系统在对抗对抗攻击时表现出脆弱性,因此对其鲁棒性进行复杂的评估是非常重要的。然而,仅根据已知的攻击在最坏情况下评估鲁棒性是不全面的,更不用说其中一些攻击在现实世界中甚至很少发生。此外,安全关键数据的分布通常是多模态的,而大多数传统的攻击和评估方法都集中在单一模态上。为了解决上述挑战,我们提出了一个基于流的多模态安全关键场景生成器评估决策算法。建议的生成模型进行了优化与加权似然最大化和基于梯度的采样过程中,集成,以提高采样效率。通过查询任务算法生成安全关键场景,并且生成的场景的对数似然与风险水平成比例。自动驾驶任务的实验证明了我们在测试效率和多模态建模能力方面的优势。我们评估六个强化学习算法与我们生成的交通场景,并提供经验的结论,他们的鲁棒性。

英文摘要:

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.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑