2000 character limit reached
Learning temporal formulas from examples is hard
Published 26 Dec 2023 in cs.LG, cs.AI, cs.FL, and cs.LO | (2312.16336v1)
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.
- Uncertainty-aware signal temporal logic inference.Ā In Bloem, R., Dimitrova, R., Fan, C., &Ā Sharygina, N. (Eds.), International Conference on Software Verification, VSTTE, Vol.Ā 13124 of Lecture Notes in Computer Science, pp.Ā 61ā85. Springer.
- A Decision Tree Approach to Data Classification using Signal Temporal Logic.Ā In Hybrid Systems: Computation and Control, HSCC.
- Learning interpretable models expressed in linear temporal logic.Ā International Conference on Automated Planning and Scheduling, ICAPS, 29.
- Analytical approach to parallel repetition.Ā In Symposium on Theory of Computing, STOC, pp.Ā 624ā633.
- Learning properties in LTL ā©\capā© ACTL from positive examples only.Ā In Formal Methods in Computer Aided Design, FMCAD.
- The complexity of learning linear temporal formulas from examples.Ā In Chandlee, J., Eyraud, R., Heinz, J., Jardine, A., &Ā van Zaanen, M. (Eds.), International Conference on Grammatical Inference, ICGI, Vol.Ā 153 of Proceedings of Machine Learning Research, pp.Ā 237ā250. PMLR.
- Maxsat-based temporal logic inference from noisy data.Ā Innovations in Systems and Software Engineering, 18(3), 427ā442.
- Gold, E.Ā M. (1978). Complexity of automaton identification from given data.Ā Information and Control, 37(3), 302ā320.
- Bayesian inference of linear temporal logic specifications for contrastive explanations.Ā In International Joint Conference on Artificial Intelligence, IJCAI.
- General LTL specification mining.Ā In International Conference on Automated Software Engineering, (ASE).
- Specification sketching for linear temporal logic.Ā In AndrĆ©, Ć.,Ā &Ā Sun, J. (Eds.), International Symposium on Automated Technology for Verification and Analysis, ATVA, Vol.Ā 14216 of Lecture Notes in Computer Science, pp.Ā 26ā48. Springer.
- Learning linear temporal properties.Ā In Formal Methods in Computer Aided Design, FMCAD, pp.Ā 1ā10.
- The minimum consistent DFA problem cannot be approximated within any polynomial.Ā Journal of the ACM, 40(1), 95ā142.
- Pnueli, A. (1977). The temporal logic of programs.Ā In Symposium on Foundations of Computer Science, SFCS.
- Scalable anytime algorithms for learning fragments of linear temporal logic.Ā In Fisman, D.,Ā &Ā Rosu, G. (Eds.), International Conference on Tools and Algorithms for the Construction and Analysis of Systems, TACAS, Vol.Ā 13243 of Lecture Notes in Computer Science, pp.Ā 263ā280. Springer.
- Synthesizing efficiently monitorable formulas in metric temporal logic.Ā In International Conference on Verification, Model Checking, and Abstract Interpretation, VMCAI.
- Learning interpretable models in the property specification language.Ā In International Joint Conference on Artificial Intelligence, IJCAI.
Paper Prompts
Sign up for free to create and run prompts on this paper using GPT-5.
Top Community Prompts
Collections
Sign up for free to add this paper to one or more collections.