Multi-Task Learning In Conditional Random Fields For Chunking In Shallow Semantic Parsing

In the field of natural language processing, semantic analysis is an indispensable phase. Through semantic role labeling, shallow semantic information could be obtained, which corresponds to deep semantic concept. Semantic Role Labeling bridges the gap between grammar and semantics.Preprocessing need to be conducted for semantic role labeling tasks, including:word segmentation, POS tagging, chunk analysis, etc. among which chunk analysis bears direct relation to deep semantic information, porting great influence to the quality of semantic analysis. Thus, it has attracts broad attention among NLP researchers.Traditional shallow semantic parsing systems for chunk analysis are implemented in statistic model with Single-task Learning (STL) mode, such as Conditional Random Files Model, Support Vector Machines, Maximum Entropy Markov Models. Although promising results have been obtained, yet, we usually have no idea of what a good model is like, neither do we know which features to select.Fortunately, Multi-task Learning (MTL) provides us with a substantial solution to work out this nontrivial problem. Multi-task Learning is an approach to inductive transfer that emphasizes learning multiple tasks in parallel while using a shared representation, so that what is learned by all tasks is available to the target task.Alternating Structure Optimization (ASO) is a recently proposed linear Multitask Learning algorithm. Its effective has been verified in both semi-supervised as well as supervised methods, our experiemts on Chinese Tree Bank 2.0 data set demonstrate that ASO is also useful for Chinese text. Since previous methods all necessitate external resource as a prerequisitem, the feasibility of employing ASO to further improve the performance merely rests on the labeled data on hand proves to be a task deserving close scrutiny. Catering to this challenging while untapped problem, this paper presents a novel application of ASO to the subtask of Shallow Semantic Parsing:Chunking. Our experiments on Chinese Treebank 5.0 present promising result in chunk analysis, and the error rate is reduced by 5.72%, proposing a profound way to further improve the performance.