An Efficient Algorithm for the Inference of Circuit-Free Automata
In this paper, a recently introduced grammatical inference method is reviewed. In this method, a non left(right)-recursive regular grammar is built in an incremental way: as each training sample is presented, it is parsed by the current (error-correcting extended) grammar, minimizing explicitly, by dynamic programming, the number of error-rules needed. These error-rules are then added to the grammar. This procedure has proved to be well suited for capturing the relevant information associated with the lengths of the substructures of the patterns to be analized, and with their concatenation. A stochastic extension of the method is presented, and some alternative approaches for estimating the probabilities of both the error and non-error rules are discussed. Finally, the results of some experiments with speech samples, which show the capabilities of the proposed method, are summarized.
KeywordsAutomatic Speech Recognition Attribute Grammar Grammatical Inference Error Rule Regular Grammar
Unable to display preview. Download preview PDF.
- (DeMori, 8l).R. De Mori: “Computer Models of Speech using Fuzzy Algorithms”, Plenum Press, New York, 1981.Google Scholar
- (Feldman, 72).
- (Forney, 73).
- (Fu, 82).K.S. Fu: “Syntactic Pattern Recognition and Applications”. Ed. Prentice-Hall. 1982.Google Scholar
- (Fu, 83).
- (Hunt, 75).
- (Itoga, 8l).S.Y. Itoga: “A New Heuristic for Inferring Regular Grammars”. IEEE-Trans, on Pattern Analysis ana Machine Intelligence, Vol. PAMI-3, N.2, 1981.Google Scholar
- (Jelinek, 76).
- (Juang, 85).B.H.Juang, L.R. Rabiner, S.E. Levinson, M.M. Sondhi: “Recent developments in the application of Hidden Markov Models to Speaker-Independent Isolated Word Recognition”. IEEE-ICASSP; pp. 9–12, 1985.Google Scholar
- (Miclet, 80).
- (Richetin, 84).
- (Rulot, 86).H.Rulot, E.Vidal: “Modelling (sub)string-length-based constraints through a Grammatical Inference method”. NATO-ASI on Pattern Recognition and applications. Spa (Belgium). June, 1986.Google Scholar
- (Rusell, 85).M.J.Rusell, R.K.Moore: “Explicit modelling of state occupancy in Hidden Markov Models for automatic speech recognition”. IEEE-ICASSP, pp. 5–8, 1985.Google Scholar
- (Thomason, 86).
- (Vernadat, 84).F. Vernadat, M. Richetin: “Regular Inference for syntactic pattern recognition: a case study”. IEEE-ICPR, pp. 1370–1372, 1984.Google Scholar
- (Vidal, 85).E.Vidal, F.Casacuberta, E.Sanchís, J.Benedí: “A General Fuzzy-Parsing Scheme for Speech Recognition”. NATO-ASI New Systems and Architectures for Automatic Speech Recognition and Synthesis. R.DeMori, C.Y.Suen (eds.). Springer-Verlag. pp. 427–446, 1985.Google Scholar