Automatic Inference of Minimalist Grammars using an SMT-Solver
This work addresses the challenge of grammar inference for computational linguistics, offering an incremental improvement by automating the process with SMT-solvers and cost-based evaluation.
The authors tackled the problem of automatically inferring Minimalist Grammars from annotated sentences by introducing a novel parser encoded as first-order logic formulae for SMT-solvers and a procedure using it, resulting in inferred grammars that aligned with contemporary syntax theories when optimized with cost functions based on Minimum Description Length and Subset principles.
We introduce (1) a novel parser for Minimalist Grammars (MG), encoded as a system of first-order logic formulae that may be evaluated using an SMT-solver, and (2) a novel procedure for inferring Minimalist Grammars using this parser. The input to this procedure is a sequence of sentences that have been annotated with syntactic relations such as semantic role labels (connecting arguments to predicates) and subject-verb agreement. The output of this procedure is a set of minimalist grammars, each of which is able to parse the sentences in the input sequence such that the parse for a sentence has the same syntactic relations as those specified in the annotation for that sentence. We applied this procedure to a set of sentences annotated with syntactic relations and evaluated the inferred grammars using cost functions inspired by the Minimum Description Length principle and the Subset principle. Inferred grammars that were optimal with respect to certain combinations of these cost functions were found to align with contemporary theories of syntax.