Learning Tree Automata with Term Rewriting
For researchers in grammatical inference and tree automata learning, this work offers an incremental method to reduce query complexity by incorporating domain knowledge via term rewriting.
The paper extends Angluin-style learning for tree automata by integrating term rewriting to infer query answers, reducing query complexity. Examples show significant reduction in queries for natural properties.
We present an extension of the Angluin-style learning algorithm for tree automata that incorporates deductive inference. The learning algorithm is provided with a term rewriting system that specifies properties of the target tree language (e.g., the order of subtrees under a symbol f is irrelevant). This term rewriting system is used to infer answers to some queries, which reduces the query complexity of the learning algorithm. We present examples of rewrite systems that express natural properties of tree-structured data, which yield a significant reduction in the number of queries.