Title: Reducing Grounded Learning Tasks To Grammatical Inference 

Benjamin Boerschinger (with Bevan K. Jones and Mark Johnson)

Abstract:

It is often assumed that grounded learning tasks are beyond the scope of 
grammatical inference techniques. We show that the grounded task of learning a
semantic parser from ambiguous training data can be reduced to a Probabilistic
Context-Free Grammar learning task in a way that gives state of the art results.
We further show that additionally letting our model learn the languages
canonical word order improves its performance. While our model is surprisingly
simple, it has certain obvious shortcomings which we will point out, and we
will sketch how we plan to address these problems in future research.