Learning Residual Alternating Automata

Sebastian Berndt, Maciej Liskiewicz, Matthias Lutter, Rüdiger Reischuk

Abstract

Residuality plays an essential role for learning finite automata. While residual deterministic and non-deterministic automata have been understood quite well, fundamental questions concerning alternating automata (AFA) remain open. Recently, Angluin, Eisenstat, and Fisman (2015) have initiated a systematic study of residual AFAs and proposed an algorithm called AL∗-an extension of the popular L∗ algorithm-to learn AFAs. Based on computer experiments they have conjectured that AL∗ produces residual AFAs, but have not been able to give a proof. In this paper we disprove this conjecture by constructing a counterexample. As our main positive result we design an efficient learning algorithm, named AL∗∗, and give a proof that it outputs residual AFAs only. In addition, we investigate the succinctness of these different FA types in more detail.

OriginalspracheEnglisch
TitelProc. 31st AAAI Conference on Artificial Intelligence (AAAI 2017)
Redakteure/-innenSatinder Singh
Seitenumfang26
Herausgeber (Verlag) AAAI Press
Erscheinungsdatum08.03.2017
Seiten1749-1755
ISBN (Print)9781577357858
DOIs
PublikationsstatusVeröffentlicht - 08.03.2017
VeranstaltungThirty-First AAAI Conference on Artificial Intelligence
- San Francisco, USA / Vereinigte Staaten
Dauer: 04.02.201709.02.2017
https://aaai.org/Conferences/AAAI/aaai17.php

Fingerprint

Untersuchen Sie die Forschungsthemen von „Learning Residual Alternating Automata“. Zusammen bilden sie einen einzigartigen Fingerprint.

Zitieren