Query Learning Algorithm for Residual Symbolic Finite Automata
This work addresses the challenge of learning automata with succinct representations for large alphabets, which is incremental as it builds on existing SFA and RFA concepts.
The authors tackled the problem of learning residual symbolic finite automata (RSFAs) by proposing a query learning algorithm that efficiently handles huge alphabets and outperforms existing deterministic SFA learning methods, demonstrating that non-determinism provides greater efficiency gains in symbolic automata learning than in non-symbolic cases.
We propose a query learning algorithm for residual symbolic finite automata (RSFAs). Symbolic finite automata (SFAs) are finite automata whose transitions are labeled by predicates over a Boolean algebra, in which a big collection of characters leading the same transition may be represented by a single predicate. Residual finite automata (RFAs) are a special type of non-deterministic finite automata which can be exponentially smaller than the minimum deterministic finite automata and have a favorable property for learning algorithms. RSFAs have both properties of SFAs and RFAs and can have more succinct representation of transitions and fewer states than RFAs and deterministic SFAs accepting the same language. The implementation of our algorithm efficiently learns RSFAs over a huge alphabet and outperforms an existing learning algorithm for deterministic SFAs. The result also shows that the benefit of non-determinism in efficiency is even larger in learning SFAs than non-symbolic automata.