Improving neural network verification through spurious region guided refinement

• Main Authors: YANG, Pengfei, LI, Renjue, LI, Jianlin, HUANG, Cheng Chao, WANG, Jingyi, SUN, Jun, XUE, Bai, ZHANG, Lijun
• 格式: text
• 語言: English
• 出版: Institutional Knowledge at Singapore Management University 2021
• 主題: deep neural networks; spurious regions; DeepPoly; Software Engineering
• 在線閱讀: https://ink.library.smu.edu.sg/sis_research/6057; https://ink.library.smu.edu.sg/context/sis_research/article/7060/viewcontent/Yang2021_Chapter_ImprovingNeuralNetworkVerifica.pdf
• 機構: Singapore Management University

We propose a spurious region guided refinement approach for robustness verification of deep neural networks. Our method starts with applying the DeepPoly abstract domain to analyze the network. If the robustness property cannot be verified, the result is inconclusive. Due to the over-approximation, the computed region in the abstraction may be spurious in the sense that it does not contain any true counterexample. Our goal is to identify such spurious regions and use them to guide the abstraction refinement. The core idea is to make use of the obtained constraints of the abstraction to infer new bounds for the neurons. This is achieved by linear programming techniques. With the new bounds, we iteratively apply DeepPoly, aiming to eliminate spurious regions. We have implemented our approach in a prototypical tool DeepSRGR. Experimental results show that a large amount of regions can be identified as spurious, and as a result, the precision of DeepPoly can be significantly improved. As a side contribution, we show that our approach can be applied to verify quantitative robustness properties.