We proposes an algorithm to learn automata innite alphabets, or at least too large to enumerate. We apply it to dene a generic model intended for regression, with transitions constrained by intervals over the alphabet. The algorithm is based on the Red & Blue framework for learning from an input sample. We show two small case studies where the alphabets are respectively the natural and real numbers, and show how nice properties of automata models like interpretability and graphical representation transfer to regression where typical models are hard to interpret.
|Title of host publication||Proceedings of The 13th International Conference on Grammatical Inference|
|Subtitle of host publication||The JMLR Workshop and Conference, The Sequence PredictIction ChallengE (SPiCe)|
|Editors||S. Verwer, M. van Zaanen, R. Smetsers|
|Number of pages||5|
|Publication status||Published - 2016|
|Event||13th International Conference on Grammatical Inference, ICGI 2016 - Delft, Netherlands|
Duration: 5 Oct 2016 → 7 Oct 2016
|Name||JMLR: Workshop and Conference Proceedings|
|Conference||13th International Conference on Grammatical Inference, ICGI 2016|
|Period||5/10/16 → 7/10/16|
- Passive Learning
- Deterministic Finite Automata
Pellegrino, N., Hammerschmidt, C., Verwer, S., & Lin, Q. (2016). Learning Deterministic Finite Automata from Innite Alphabets. In S. Verwer, M. van Zaanen, & R. Smetsers (Eds.), Proceedings of The 13th International Conference on Grammatical Inference: The JMLR Workshop and Conference, The Sequence PredictIction ChallengE (SPiCe) (Vol. 57, pp. 69-72). (JMLR: Workshop and Conference Proceedings). JMLR.