| dc.creator | Ghahramani, Zoubin | |
| dc.creator | Jordan, Michael I. | |
| dc.date | 2004-10-20T20:49:14Z | |
| dc.date | 2004-10-20T20:49:14Z | |
| dc.date | 1996-02-09 | |
| dc.date.accessioned | 2013-10-09T02:48:31Z | |
| dc.date.available | 2013-10-09T02:48:31Z | |
| dc.date.issued | 2013-10-09 | |
| dc.identifier | AIM-1561 | |
| dc.identifier | CBCL-130 | |
| dc.identifier | http://hdl.handle.net/1721.1/7188 | |
| dc.identifier.uri | http://koha.mediu.edu.my:8181/xmlui/handle/1721 | |
| dc.description | We present a framework for learning in hidden Markov models with distributed state representations. Within this framework, we derive a learning algorithm based on the Expectation--Maximization (EM) procedure for maximum likelihood estimation. Analogous to the standard Baum-Welch update rules, the M-step of our algorithm is exact and can be solved analytically. However, due to the combinatorial nature of the hidden state representation, the exact E-step is intractable. A simple and tractable mean field approximation is derived. Empirical results on a set of problems suggest that both the mean field approximation and Gibbs sampling are viable alternatives to the computationally expensive exact algorithm. | |
| dc.format | 7 p. | |
| dc.format | 198365 bytes | |
| dc.format | 244196 bytes | |
| dc.format | application/postscript | |
| dc.format | application/pdf | |
| dc.language | en_US | |
| dc.relation | AIM-1561 | |
| dc.relation | CBCL-130 | |
| dc.subject | AI | |
| dc.subject | MIT | |
| dc.subject | Artificial Intelligence | |
| dc.subject | Hidden Markov Models | |
| dc.subject | sNeural networks | |
| dc.subject | Time series | |
| dc.subject | Mean field theory | |
| dc.subject | Gibbs sampling | |
| dc.subject | sFactorial | |
| dc.subject | Learning algorithms | |
| dc.subject | Machine learning | |
| dc.title | Factorial Hidden Markov Models |
| Files | Size | Format | View |
|---|---|---|---|
|
There are no files associated with this item. |
|||