Incremental Learning of Context Free Grammars by Bridging Rule Generation and Search for Semi-optimum Rule Sets

Abstract

This paper describes novel methods of learning general context free grammars from sample strings, which are implemented in Synapse system. Main features of the system are incremental learning, rule generation based on bottom-up parsing of positive samples, and search for rule sets. From the results of parsing, a rule generation process, called “bridging,” synthesizes the production rules that make up any lacking parts of an incomplete derivation tree for each positive string. To solve the fundamental problem of complexity for learning CFG, we employ methods of searching for non-minimum, semi-optimum sets of rules as well as incremental learning based on related grammars. One of the methods is search strategy called “serial search,” which finds additional rules for each positive sample and not to find the minimum rule set for all positive samples as in global search. The other methods are not to minimize nonterminal symbols in rule generation and to restrict the form of generated rules. The paper shows experimental results and compares various synthesis methods.