Paper Info
Title
Initial evaluation of hidden dynamic models on conversational speech.
Abstract
Conversational speech recognition is a challenging problem primarily because speakers rarely fully articulate sounds. A successful speech recognition approach must infer intended spectral targets from the speech data, or develop a method of dealing with large variances in the data. Hidden dynamic models (HDMs) attempt to automatically learn such targets in a hidden feature space using models that integrate linguistic information with constrained temporal trajectory models. HDMs are a radical departure from conventional hidden Markov models (HMMs), which simply account for variation in the observed data. We present an initial evaluation of such models on a conversational speech recognition task involving a subset of the SWITCHBOARD corpus. We show that in an N-best rescoring paradigm, HDMs are capable of delivering performance competitive with HMMs.
Year
DOI
Venue
1999
10.1109/ICASSP.1999.758074
ICASSP
Keywords
Field
DocType
hidden feature space,conventional hidden markov model,successful speech recognition approach,conversational speech recognition,speech data,hidden dynamic model,switchboard corpus,conversational speech recognition task,observed data,n-best rescoring paradigm,initial evaluation,low pass filters,performance,linguistic information,speech processing,natural languages,speech recognition,acoustical engineering,hidden markov model,trajectory,hidden markov models,feature space,loudspeakers,hmm
Rule-based machine translation,Feature vector,Pattern recognition,Computer science,Speech recognition,Dynamic models,Artificial intelligence,Natural language processing,Hidden Markov model,Trajectory
Conference
ISSN
ISBN
Citations 
1520-6149
0-7803-5041-3
21
PageRank 
References 
Authors
2.96
2
9
Name
Order
Citations
PageRank
J. Picone120226.15
S. Pike2212.96
R. Regan3212.96
T. Kamm4303.65
J. Bridle5178107.11
Deng, Li69691728.14
Jeff Z. Ma713315.79
H. Richards8212.96
Mike Schuster92303111.71