Learning Linear Temporal Logic formulas from Examples

We are interested in constructing linear temporal formulas (LTL) from finite traces. The problem has attracted a lot of attention recently in different communities: software engineering, verification and control of cyber-physical systems, and artificial intelligence. In this talk, I will report on two papers: - a complexity-theoretic study, published in ICGI'21 with Guillaume Lagarde - a tool, published in TACAS'22 with Ritam Raha, Rajarshi Roy, and Daniel Neider The goal of these papers is to develop new algorithms for learning fragments of LTL by analysing formulas at a syntactic level.

Authors

• Nathanael Fijalkow

Keywords

No keywords are indexed for this paper.