The hierarchical hidden Markov model: Analysis and applications

被引:498
作者
Fine, S [1 ]
Singer, Y
Tishby, N
机构
[1] Hebrew Univ Jerusalem, Inst Comp Sci, IL-91904 Jerusalem, Israel
[2] AT&T Bell Labs, Florham Pk, NJ 07932 USA
关键词
statistical models; temporal pattern recognition; hidden variable models; cursive handwriting;
D O I
10.1023/A:1007469218079
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
We introduce, analyze and demonstrate a recursive hierarchical generalization of the widely used hidden Markov models, which we name Hierarchical Hidden Markov Models (HHMM). Our model is motivated by the complex multi-scale structure which appears in many natural sequences, particularly in language, handwriting and speech. We seek a systematic unsupervised approach to the modeling of such structures. By extending the standard Baum-Welch (forward-backward) algorithm, we derive an efficient procedure for estimating the model parameters from unlabeled data. We then use the trained model for automatic hierarchical parsing of observation sequences. We describe two applications of our model and its parameter estimation procedure. In the first application we show how to construct hierarchical models of natural English text. In these models different levels of the hierarchy correspond to structures on different length scales in the text. In the second application we demonstrate how HHMMs can be used to automatically identify repeated strokes that represent combination of letters in cursive handwriting.
引用
收藏
页码:41 / 62
页数:22
相关论文
共 23 条
[1]  
ABE N, 1992, MACH LEARN, V9, P205, DOI 10.1007/BF00992677
[2]  
[Anonymous], IEEE ASSP MAGAZINE
[3]  
[Anonymous], 1989, P IEEE
[4]   HIDDEN MARKOV-MODELS OF BIOLOGICAL PRIMARY SEQUENCE INFORMATION [J].
BALDI, P ;
CHAUVIN, Y ;
HUNKAPILLER, T ;
MCCLURE, MA .
PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES OF THE UNITED STATES OF AMERICA, 1994, 91 (03) :1059-1063
[5]   STATISTICAL INFERENCE FOR PROBABILISTIC FUNCTIONS OF FINITE STATE MARKOV CHAINS [J].
BAUM, LE ;
PETRIE, T .
ANNALS OF MATHEMATICAL STATISTICS, 1966, 37 (06) :1554-&
[6]  
Bengio Y., 1995, Advances in Neural Information Processing Systems 7, P427
[7]  
Cover T. M., 2005, ELEM INF THEORY, DOI 10.1002/047174882X
[8]   MAXIMUM LIKELIHOOD FROM INCOMPLETE DATA VIA EM ALGORITHM [J].
DEMPSTER, AP ;
LAIRD, NM ;
RUBIN, DB .
JOURNAL OF THE ROYAL STATISTICAL SOCIETY SERIES B-METHODOLOGICAL, 1977, 39 (01) :1-38
[9]   Hidden Markov modelling of simultaneously recorded cells in the associative cortex of behaving monkeys [J].
Gat, I ;
Tishby, N ;
Abeles, M .
NETWORK-COMPUTATION IN NEURAL SYSTEMS, 1997, 8 (03) :297-322
[10]   Factorial hidden Markov models [J].
Ghahramani, Z ;
Jordan, MI .
MACHINE LEARNING, 1997, 29 (2-3) :245-273