Learning temporal formulas from examples is hard
Abstract
We study the problem of learning linear temporal logic (LTL) formulas from examples, as a first step towards expressing a property separating positive and negative instances in a way that is comprehensible for humans. In this paper we initiate the study of the computational complexity of the problem. Our main results are hardness results: we show that the LTL learning problem is NP-complete, both for the full logic and for almost all of its fragments. This motivates the search for efficient heuristics, and highlights the complexity of expressing separating properties in concise natural language.
Cite
@article{arxiv.2312.16336,
title = {Learning temporal formulas from examples is hard},
author = {Corto Mascle and Nathanaël Fijalkow and Guillaume Lagarde},
journal= {arXiv preprint arXiv:2312.16336},
year = {2023}
}
Comments
This article is a long version of the article arXiv:2102.00876 presented in the International Conference on Grammatical Inference (ICGI) in 2021. It includes much stronger and more general results than the extended abstract. Submitted to a journal