The Hierarchical Dirichlet Process Hidden Semi-Markov Model
There is much interest in the Hierarchical Dirichlet Process Hidden Markov Model (HDP-HMM) as a natural Bayesian nonparametric extension of the traditional HMM. However, in many settings the HDP-HMM's strict Markovian constraints are undesirable, particularly if we wish to learn or encode non-g...
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Article |
| Language: | en_US |
| Published: |
Association for Uncertainty in Artificial Intelligence (AUAI)
2013
|
| Online Access: | http://hdl.handle.net/1721.1/79638 https://orcid.org/0000-0003-0149-5888 |
| _version_ | 1811083907575054336 |
|---|---|
| author | Johnson, Matthew James Willsky, Alan S |
| author2 | Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science |
| author_facet | Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science Johnson, Matthew James Willsky, Alan S |
| author_sort | Johnson, Matthew James |
| collection | MIT |
| description | There is much interest in the Hierarchical Dirichlet Process Hidden Markov Model (HDP-HMM) as a natural Bayesian nonparametric extension of the traditional HMM. However, in many settings the HDP-HMM's strict Markovian constraints are undesirable, particularly if we wish to learn or encode non-geometric state durations. We can extend the HDP-HMM to capture such structure by drawing upon explicit-duration semi- Markovianity, which has been developed in the parametric setting to allow construction of highly interpretable models that admit natural prior information on state durations. In this paper we introduce the explicitduration HDP-HSMM and develop posterior sampling algorithms for efficient inference in both the direct-assignment and weak-limit approximation settings. We demonstrate the utility of the model and our inference methods on synthetic data as well as experiments on a speaker diarization problem and an example of learning the patterns in Morse code. |
| first_indexed | 2024-09-23T12:41:31Z |
| format | Article |
| id | mit-1721.1/79638 |
| institution | Massachusetts Institute of Technology |
| language | en_US |
| last_indexed | 2024-09-23T12:41:31Z |
| publishDate | 2013 |
| publisher | Association for Uncertainty in Artificial Intelligence (AUAI) |
| record_format | dspace |
| spelling | mit-1721.1/796382022-09-28T09:25:29Z The Hierarchical Dirichlet Process Hidden Semi-Markov Model Johnson, Matthew James Willsky, Alan S Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science Johnson, Matthew James Willsky, Alan S. There is much interest in the Hierarchical Dirichlet Process Hidden Markov Model (HDP-HMM) as a natural Bayesian nonparametric extension of the traditional HMM. However, in many settings the HDP-HMM's strict Markovian constraints are undesirable, particularly if we wish to learn or encode non-geometric state durations. We can extend the HDP-HMM to capture such structure by drawing upon explicit-duration semi- Markovianity, which has been developed in the parametric setting to allow construction of highly interpretable models that admit natural prior information on state durations. In this paper we introduce the explicitduration HDP-HSMM and develop posterior sampling algorithms for efficient inference in both the direct-assignment and weak-limit approximation settings. We demonstrate the utility of the model and our inference methods on synthetic data as well as experiments on a speaker diarization problem and an example of learning the patterns in Morse code. 2013-07-22T14:01:16Z 2013-07-22T14:01:16Z 2010 Article http://purl.org/eprint/type/ConferencePaper http://hdl.handle.net/1721.1/79638 Matthew Johnson and Alan Willsky. "The Hierarchical Dirichlet Process Hidden Semi-Markov Model", Proceedings of the Twenty-Sixth Conference Annual Conference on Uncertainty in Artificial Intelligence (UAI-10), AUAI Press (2010): 252-259. https://orcid.org/0000-0003-0149-5888 en_US http://event.cwi.nl/uai2010/papers/UAI2010_0193.pdf Proceedings of the Twenty-Sixth Conference on Uncertainty in Artificial Intelligence (UAI2010) Creative Commons Attribution-Noncommercial-Share Alike 3.0 http://creativecommons.org/licenses/by-nc-sa/3.0/ application/pdf Association for Uncertainty in Artificial Intelligence (AUAI) Willsky via Amy Stout |
| spellingShingle | Johnson, Matthew James Willsky, Alan S The Hierarchical Dirichlet Process Hidden Semi-Markov Model |
| title | The Hierarchical Dirichlet Process Hidden Semi-Markov Model |
| title_full | The Hierarchical Dirichlet Process Hidden Semi-Markov Model |
| title_fullStr | The Hierarchical Dirichlet Process Hidden Semi-Markov Model |
| title_full_unstemmed | The Hierarchical Dirichlet Process Hidden Semi-Markov Model |
| title_short | The Hierarchical Dirichlet Process Hidden Semi-Markov Model |
| title_sort | hierarchical dirichlet process hidden semi markov model |
| url | http://hdl.handle.net/1721.1/79638 https://orcid.org/0000-0003-0149-5888 |
| work_keys_str_mv | AT johnsonmatthewjames thehierarchicaldirichletprocesshiddensemimarkovmodel AT willskyalans thehierarchicaldirichletprocesshiddensemimarkovmodel AT johnsonmatthewjames hierarchicaldirichletprocesshiddensemimarkovmodel AT willskyalans hierarchicaldirichletprocesshiddensemimarkovmodel |