TY - CHAP
T1 - A Literature Review on Verification and Abstraction of Neural Networks Within the Formal Methods Community
AU - Kanav, Sudeep
AU - Křetínský, Jan
AU - Rieder, Sabine
N1 - Publisher Copyright:
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.
PY - 2025
Y1 - 2025
N2 - With the increasing interest in applying neural networks (NNs) to safety-critical problems like autonomous driving or unmanned aircrafts, ensuring the reliability of these NNs becomes essential. Therefore, several verification techniques have been proposed in recent years. Additionally, various abstraction techniques have been developed to enable verification of larger NNs. As the area develops, different surveys on verification of NNs are being published. However, we are missing a systematic summarization of knowledge through the lens of formal methods. In this literature review, we provide a systematic overview of techniques for verification and abstraction of NNs published in well-known formal verification conferences during the last ten years.
AB - With the increasing interest in applying neural networks (NNs) to safety-critical problems like autonomous driving or unmanned aircrafts, ensuring the reliability of these NNs becomes essential. Therefore, several verification techniques have been proposed in recent years. Additionally, various abstraction techniques have been developed to enable verification of larger NNs. As the area develops, different surveys on verification of NNs are being published. However, we are missing a systematic summarization of knowledge through the lens of formal methods. In this literature review, we provide a systematic overview of techniques for verification and abstraction of NNs published in well-known formal verification conferences during the last ten years.
UR - http://www.scopus.com/inward/record.url?scp=85212145354&partnerID=8YFLogxK
U2 - 10.1007/978-3-031-75778-5_3
DO - 10.1007/978-3-031-75778-5_3
M3 - Chapter
AN - SCOPUS:85212145354
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 39
EP - 65
BT - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
PB - Springer Science and Business Media Deutschland GmbH
ER -