The Acquisition of Word Order by a Computational Learning System

In: Proceedings of CoNLL-2000 and LLL-2000, pages 209-218, Lisbon, Portugal, 2000. The Acquisition of Word Order by a Computational Learning System Aline Villavicencio Computer Laboratory, University of Cambridge New Museums Site, Cambridge, CB2 3QG, England, UK Aline. Villavicencio@cl. cam. ac. uk Abstract The purpose of this work is to investigate the process of grammatical acquisition from data. We are using a computational learning sys- tern that is composed of a Universal Grammar with associated parameters, and a learning al- gorithm, following the Principles and Parame- ters Theory. The Universal Grammar is imple- mented as a Unification-Based Generalised Cat- egorial Grammar, embedded in a default inher- itance network of lexical types. The learning al- gorithm receives input from a corpus annotated with logical forms and sets the parameters based on this input. This framework is used as basis to investigate several aspects of language acqui- sition. In this paper we are concentrating on the acquisition of word order for different learners. The results obtained show the different learners having a similar performance and converging to- wards the target grammar given the input data available, regardless of their starting points. It also shows how the amount of noise present in the input data affects the speed of convergence of the learners towards the target. 1 Introduction In trying to solve the question of how to get a machine to automatically learn linguistic infor- mation from data, we can look at the way people do it. Gold (1967) when investigating language identification in the limit, obtained results that implied that natural languages could not be learned only on the basis of positive evidence. These results were used as a confirmation for the proposal that children must have some innate knowledge about language, the Universal Gram- mar (UG), to help them overcome the prob- lem of the poverty of the stimulus and acquire a grammar on the basis of positive evidence only. According to Chomsky's Principles and Parameters Theory (Chomsky 1981), the UG is composed of principles and parameters, and the process of learning a language is regarded as the setting of values of a number of parame- ters, given exposure to this particular language. We employ this idea in the learning framework implemented. In this work we are interested in investigating the acquisition of grammatical knowledge from data, focusing on the acquisition of word or- der, that reflects the underlying order in which constituents occur in different languages (e.g. SVO and SOV languages). The learning sys- tem is equipped with a UG and associated pa- rameters, encoded as a Unification-Based Gen- eralised Categorial Grammar, and a learning al- gorithm that fixes the values of the parameters to a particular language. The learning algo- rithm follows the Bayesian Incremental Param- eter Setting (BIPS) algorithm (Briscoe 1999), and when setting the parameters it uses a Mini- mum Description Length (MDL) style bias to choose the most probable grammar that de- scribes the data well, given the goal of converg- ing to the target grammar. In …

Content truncated to 500 words