We propose a novel application of selfattention networks towards grammar induction. We present an attention-based supertagger for a refined type-logical grammar, trained on constructing types inductively. In addition to achieving a high overall type accuracy, our model is able to learn the syntax of the grammar's type system along with its denotational semantics. This lifts the closed world assumption commonly made by lexicalized grammar supertaggers, greatly enhancing its generalization potential. This is evidenced both by its adequate accuracy over sparse word types and its ability to correctly construct complex types never seen during training, which, to the best of our knowledge, was as of yet unaccomplished.

Incremental parsers have potential advantages for applications like language modeling for machine translation and speech recognition. We describe a new algorithm for incremental transition-based Combinatory Categorial Grammar parsing. As English CCGbank derivations are mostly right branching and non-incremental, we design our algorithm based on the dependencies resolved rather than the derivation. We introduce two new actions in the shift-reduce paradigm based on the idea of 'revealing' (Pareschi and Steedman, 1987) the required information during parsing. On the standard CCGbank test data, our algorithm achieved improvements of 0.88% in labeled and 2.0% in unlabeled F-score over a greedy non-incremental shift-reduce parser.

No abstract

In this paper, we present an approach for automatically creating a combinatory categorial grammar (CCG) treebank from a dependency treebank for the subject-object-verb language Hindi. Rather than a direct conversion from dependency trees to CCG trees, we propose a two stage approach: a language independent generic algorithm first extracts a CCG lexicon from the dependency treebank. An exhaustive CCG parser then creates a treebank of CCG derivations. We also discuss special cases of this generic algorithm to handle linguistic phenomena specific to Hindi. In doing so we extract different constructions with long-range dependencies like coordinate constructions and non-projective dependencies resulting from constructions like relative clauses, noun elaboration and verbal modifiers.

We present a neural network based shiftreduce CCG parser, the first neural-network based parser for CCG. We also study the impact of neural network based tagging models, and greedy versus beam-search parsing, by using a structured neural network model. Our greedy parser obtains a labeled F-score of 83.27%, the best reported result for greedy CCG parsing in the literature (an improvement of 2.5% over a perceptron based greedy parser) and is more than three times faster. With a beam, our structured neural network model gives a labeled F-score of 85.57% which is 0.6% better than the perceptron based counterpart.