Algorithms and Training for Weighted Multiset Automata and Regular Expressions

TL;DR

This paper explores weighted multiset automata, detailing their construction from regular expressions, proposing training methods for weight learning, and analyzing efficient computation of inside weights.

Contribution

It introduces methods to construct weighted multiset automata from regular expressions and proposes training techniques for weight learning from data.

Findings

Methods for constructing weighted multiset automata from regular expressions

Training algorithms for learning weights from data

Efficient computation strategies for inside weights

Abstract

Multiset automata are a class of automata for which the symbols can be read in any order and obtain the same result. We investigate weighted multiset automata and show how to construct them from weighted regular expressions. We present training methods to learn the weights for weighted regular expressions and for general multiset automata from data. Finally, we examine situations in which inside weights can be computed more efficiently.