Learning language using genetic algorithms

Abstract

Strict pattern-based methods of grammar induction are often frustrated by the apparently inexhaustible variety of novel word combinations in large corpora. Statistical methods offer a possible solution by allowing frequent well-formed expressions to overwhelm the infrequent ungrammatical ones. They also have the desirable property of being able to construct robust grammars from positive instances alone. Unfortunately, the “zero-frequency” problem entails assigning a small probability to all possible word patterns, thus ungrammatical n-grams become as probable as unseen grammatical ones. Further, such grammars are unable to take advantage of inherent lexical properties that should allow infrequent words to inherit the syntactic properties of the class to which they belong.

This paper describes a genetic algorithm (GA) that adapts a population of hypothesis grammars towards a more effective model of language structure. The GA is statistically sensitive in that the utility of frequent patterns is reinforced by the persistence of efficient substructures. It also supports the view of language learning as a “bootstrapping problem,” a learning domain where it appears necessary to simultaneously discover a set of categories and a set of rules defined over them. Results from a number of tests indicate that the GA is a robust, fault-tolerant method for inferring grammars from positive examples of natural language.