Preprint available

[Yves Chauvin]

                       **DO NOT FORWARD TO OTHER GROUPS**


The following paper,

"Smooth On-Line Learning Algorithms for Hidden Markov Models"

has been placed in the neuroprose archive.
It is to be published in Neural Computation.

Further information and retrieval instructions are given below.


___________________________________________________________________________


      "Smooth On-Line Learning Algorithms for Hidden Markov Models"

                              Pierre Baldi 
                              JPL, Caltech

                              Yves Chauvin 
                              Net-ID, Inc.

A simple learning algorithm for Hidden Markov Models (HMMs) is
presented together with a number of variations. Unlike other classical 
algorithms such as the Baum-Welch algorithm, the algorithms described 
are smooth and can be used on-line (after each example presentation) 
or in batch mode, with or without the usual Viterbi most likely path 
approximation.  The simple expression of the learning algorithms
and several of their advantages result from using Boltzmann-Gibbs 
representations (normalizing exponentials) for the HMM parameters. 
All the algorithms presented are proved to be exact or approximate 
gradient optimization algorithms with respect to likelihood, 
log-likelihood or cross-entropy functions, and as such are usually
convergent. These algorithms can also be casted in the more general 
EM (Expectation-Maximization) framework where they can be viewed as
exact or approximate GEM (Generalized Expectation-Maximization) 
algorithms. The mathematical properties of the algorithms are derived 
in the appendix.

___________________________________________________________________________

Retrieval instructions:

The paper is baldi.smoothhmm.ps.Z in the neuroprose archive.

To retrieve this file from the neuroprose archives:

unix> ftp cheops.cis.ohio-state.edu
Name (cheops.cis.ohio-state.edu:becker): anonymous
Password: (use your email address)
ftp> cd pub/neuroprose
ftp> binary
ftp> get baldi.smoothhmm.ps.Z
200 PORT command successful.
150 Opening BINARY mode data connection for baldi.compbiohmm.ps.Z
.
ftp> quit
.
unix> uncompress baldi.smoothhmm.ps.Z
unix> lpr baldi.smoothhmm.ps