Formal Verification
Formal verification uses mathematical methods to rigorously prove the correctness of systems, aiming to eliminate design flaws and ensure reliable operation, especially in safety-critical applications. Current research focuses on applying formal verification to complex systems like neural networks and multi-agent systems, often leveraging techniques like abstract interpretation, interval bound propagation, and automated theorem proving, sometimes augmented by large language models for improved efficiency and automation. This field is crucial for enhancing the trustworthiness of AI systems, software, and hardware, leading to safer and more dependable technologies across various domains.
Papers
FVEL: Interactive Formal Verification Environment with Large Language Models via Theorem Proving
Xiaohan Lin, Qingxing Cao, Yinya Huang, Haiming Wang, Jianqiao Lu, Zhengying Liu, Linqi Song, Xiaodan Liang
VeriFlow: Modeling Distributions for Neural Network Verification
Faried Abu Zaid, Daniel Neider, Mustafa Yalçıner