Learning from a substructural perspective

Abstract

In this paper we study learning from a logical perspective. We show that there is a strong relationship between a learning strategy, its formal learning framework and its logical representational theory. This relationship enables one to translate learnability results from one theory to another. Moreover if we go from a classical logic theory to a substructural logic theory, we can transform learnability results of logical concepts to results for string languages. In this paper we will demonstrate such a translation by transforming the Valiant learnability result for boolean concepts to a learnability :result for a class of string pattern languages.

Computational Models of Language Acquisition

“…Finally, the measure was only tested on extremely short utterances. 1 We conjecture that it will not scale up to longer sentences (indeed, the results of Chang et al [20] indicate that the SPA measure correlates better with the specific corpus used for learning than with the learning algorithm).…”

“…The EMILE model [1,2,3] attempts to learn the grammatical structure of a language from positive examples, without prior knowledge of the grammar. It is based on the idea that expressions of the same (syntactic) type can be substituted in the same context, and hence it searches for clusters of expressions and contexts in the input, interpreting them as grammatical types.…”

A bibliographical study of grammatical inference

“…One direction has been taken by Adriaans through shallow grammars [156] (using categorial grammars (see Section 3.7.2): this theoretical work is the backbone of the EMILE prototype [157]. Another direction has been followed by Mohri [158].…”

Word Categorization Using Clustering Ensemble

“…Adriaans [9] considers language learning from the perspective of category theory. In this context, learning a categorical grammar can be seen as a reconstruction of the lexicon from examples of the language [9].…”