How Reasoning Achieves Context Integration into Syntax Parsing

Motivated by a representational model for the cross-modal interaction between language and other modalities we present a framework for the integration of contextual information into syntactic parsing. We provide a detailed description of the reasoning steps at the interface between a constraint-based parser and a semantic knowledge representation of context information. Our model implements context integration as a three-step process that involves (1) intramodal grounding, (2) cross-modal mapping and (3) semantic role inferences. We illustrate how these reasoning steps can assist in the task of syntactic disambiguation given biasing context information.

• Type: Journal
• Year: 2009
• Where: ICAI
• Authors: Patrick McCrae
• Keywords: Artificial Intelligence, Context Information, ICAI 2009, Semantic Knowledge Representation, Semantic Role Inferences