Jason Baldridge

Multi-modal combinatory categorial grammar

The paper shows how Combinatory Categorial Grammar (CCG) can be adapted to take advantage of the extra resource-sensitivity provided by the Categorial Type Logic framework. The resulting reformulation, Multi-Modal CCG, supports lexically specified control over the applicability of combinatory rules, permitting a universal rule component and shedding the need for language-specific restrictions on rules. We discuss some of the linguistic motivation for these changes, define the Multi-Modal CCG system and demonstrate how it works on some basic examples. We furthermore outline some possible extensions and address computational aspects of Multi-Modal CCG.

Coupling CCG and Hybrid Logic Dependency Semantics

Categorial grammar has traditionally used the λ-calculus to represent meaning. We present an alternative, dependency-based perspective on linguistic meaning and situate it in the computational setting. This perspective is formalized in terms of hybrid logic and has a rich yet perspicuous propositional ontology that enables a wide variety of semantic phenomena to be represented in a single meaning formalism. Finally, we show how we can couple this formalization to Combinatory Categorial Grammar to produce interpretations compositionally.

Active learning for HPSG parse selection

We describe new features and algorithms for HPSG parse selection models and address the task of creating annotated material to train them. We evaluate the ability of several sample selection methods to reduce the number of annotated sentences necessary to achieve a given level of performance. Our best method achieves a 60% reduction in the amount of training material without any loss in accuracy.

Using eye movements to determine referents in a spoken dialogue system

Most computational spoken dialogue systems take a literary approach to reference resolution. With this type of approach, entities that are mentioned by a human interactor are unified with elements in the world state based on the same principles that guide the process during text interpretation. In human-to-human interaction, however, referring is a much more collaborative process. Participants often under-specify their referents, relying on their discourse partners for feedback if more information is needed to uniquely identify a particular referent. By monitoring eye-movements during this interaction, it is possible to improve the performance of a spoken dialogue system on referring expressions that are underspecified according to the literary model. This paper describes a system currently under development that employs such a strategy.

Probabilistic Head-Driven Parsing for Discourse Structure

We describe a data-driven approach to building interpretable discourse structures for appointment scheduling dialogues. We represent discourse structures as headed trees and model them with probabilistic head-driven parsing techniques. We show that dialogue-based features regarding turn-taking and domain specific goals have a large positive impact on performance. Our best model achieves an f-score of 43.2% for labelled discourse relations and 67.9% for unlabelled ones, significantly beating a right-branching baseline that uses the most frequent relations.

Generalizing dimensionality in combinatory categorial grammar

We extend Combinatory Categorial Grammar (CCG) with a generalized notion of multidimensional sign, inspired by the types of representations found in constraint-based frameworks like HPSG or LFG. The generalized sign allows multiple levels to share information, but only in a resource-bounded way through a very restricted indexation mechanism. This improves representational perspicuity without increasing parsing complexity, in contrast to full-blown unification used in HPSG and LFG. Well-formedness of a linguistic expressions remains entirely determined by the CCG derivation. We show how the multidimensionality and perspicuity of the generalized signs lead to a simplification of previous CCG accounts of how word order and prosody can realize information structure.

Extending the Coverage of a CCG System

We demonstrate ways to enhance the coverage of a symbolic NLP system through data-intensive and machine learning techniques, while preserving the advantages of using a principled symbolic grammar formalism. We automatically acquire a large syntactic CCG lexicon from the Penn Treebank and combine it with semantic and morphological information from another hand-built lexicon using decision tree and maximum entropy classifiers. We also integrate statistical preprocessing methods in our system.