Learning Computational Grammars: A Preliminary Report

Title: Learning Computational Grammars: A Preliminary Report

Research Question: Can machine learning techniques be applied to create grammars suitable for computational use?

Method: The study focused on syntax, specifically noun phrase (NP) syntax, and used annotated data, various dependency types, and knowledge bases (grammars). The researchers applied different machine learning algorithms to three tasks: base phrase (chunk) identification, base NP recognition, and finding arbitrary noun phrases.

Results: The project produced positive results, demonstrating the potential of machine learning techniques in natural language syntax. The researchers found that the availability of annotated data, the kind of dependencies in the data, and the availability of knowledge bases (grammars) all played significant roles in the success of learning.

Implications: This study suggests that machine learning techniques can be applied to create computational grammars. It also highlights the importance of factors such as annotated data, dependency types, and knowledge bases in the success of learning. Future research should continue to explore these factors and potentially combine different machine learning techniques for more effective results.

Link to Article: https://arxiv.org/abs/0107017v1 Authors: arXiv ID: 0107017v1