Paper: Linguistic Structure As Composition And Perturbation

ACL ID P96-1044
Title Linguistic Structure As Composition And Perturbation
Venue Annual Meeting of the Association of Computational Linguistics
Session Main Conference
Year 1996

This paper discusses the problem of learn- ing language from unprocessed text and speech signals, concentrating on the prob- lem of learning a lexicon. In particular, it argues for a representation of language in which linguistic parameters like words are built by perturbing a composition of exist- ing parameters. The power of the represen- tation is demonstrated by several examples in text segmentation and compression, ac- quisition of a lexicon from raw speech, and the acquisition of mappings between text and artificial representations of meaning. 1 Motivation Language is a robust and necessarily redundant communication mechanism. Its redundancies com- monly manifest themselves as predictable patterns in speech and text signals, and it is largely these patterns that enable text and speec...