Formal Specification
Formal specification involves rigorously defining the desired behavior of systems using mathematical logic, aiming to ensure correctness and reliability. Current research focuses on automating specification creation, particularly leveraging large language models (LLMs) to translate natural language requirements into formal languages like temporal logics and refining specifications through iterative processes and machine learning techniques. This work is crucial for improving the safety and robustness of complex systems, such as AI models, robots, and software, by enabling automated verification and synthesis. The development of efficient algorithms and tools for specification generation and repair is driving progress in this field.
Papers
November 21, 2022
November 13, 2022
October 20, 2022
September 28, 2022
September 7, 2022
June 14, 2022