GRAEL: an agent-based evolutionary computing approach for natural language grammar development

This paper describes an agent-based evolutionary computing technique called GRAEL (Grammar Evolution), that is able to perform different natural language grammar optimization and induction tasks. Two different instantiations of the GRAELenvironment are described in this paper: in GRAEL1 large annotated corpora are used to bootstrap grammatical structure in a society of agents, who engage in a series of communicative attempts, during which they redistribute grammatical information to reflect useful statistics for the task of parsing. In GRAEL-2, agents are allowed to mutate grammatical information, effectively implementing grammar rule discovery in a practical context. A combination of both GRAEL-1 and GRAEL-2 can be shown to provide an interesting all-round optimization for corpus-induced grammars. gins of grammar in a computational context [Batali, 2002; Kirby, 2001J or the co-ordinated co-evolution of grammatical principles [Briscoe, 1998]. Yet so far, little or no progress has been achieved in evaluating evolutionary computing as a tool for the induction or optimization of data-driven parsing techniques. The GRAEL environment provides a suitable framework for the induction and optimization of any type of grammar for natural language in an evolutionary setting. In this paper we hope to provide an overview of GRAEL as a grammar optimization and induction technique. We wil l first outline the basic architecture of the GRAEL environment in Section 2 on the basis of a toy example. Next, we introduce GRAEL1 (Section 3) as a grammar optimization technique that can enhance corpus-induce grammars. By adding an element of mutation in GRAEL-2 we implement a method to extend the coverage of a corpus-induced grammar. We will also describe a combination of both GRAEL-1 and GRAEL-2 which can be shown to provide an interesting all-round optimization technique for corpus-induced grammars.