Towards an effective and interpretable refinement approach for DNN verification

• Main Authors: LI, Jiaying, BAI, Guangdong, PHAM, Long H., SUN, Jun
• Format: text
• Language: English
• Published: Institutional Knowledge at Singapore Management University 2023
• Subjects: abstraction debugging; abstraction refinement; neural network verification; Software Engineering
• Online Access: https://ink.library.smu.edu.sg/sis_research/8641; https://ink.library.smu.edu.sg/context/sis_research/article/9644/viewcontent/qrs23_Surgeon_av.pdf
• Institution: Singapore Management University

Recently, several abstraction refinement techniques have been proposed to improve the verification precision for deep neural networks (DNNs). However, these techniques usually take many refinement steps to verify a property and the refinement decision in each step is hard to interpret, thus hinderin...