| Presentation | 2006/12/15 Utilizing Bayesian Network and Junction Tree Decomposition for Incorporating Additional Knowledge Sources into a Statistical Acoustic Model Sakriani SAKTI, Konstantin MARKOV, Satoshi NAKAMURA, |
|---|---|
| PDF Download Page | PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | We propose a new method of incorporating the additional knowledge sources into HMM based statistical acoustic model by utilizing Bayesian network and junction tree decomposition. The aim of using such graphical model framework are: (1) to define the probabilistic relationship between information sources, (2) to find ways to decompose a global calculation on a joint probability function into a linked set of local computations so that a simplified form of the model can be reliably estimated using the available training data. We apply this framework to the problem of incorporating wide-phonetic knowledge information, which often suffers from data sparsity and memory constraints. Furthermore, we also attempt to integrate another additional knowledge, such accent and gender information. The performance of the proposed model was evaluated using LVCSR task with two different types of accented English speech data. Experimental results show that our method improves the word accuracy with respect to the standard HMM when no additional knowledge sources were used. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | Acoustic modeling / knowledge incorporation / Bayesian network / junction tree / wide-context dependency |
| Paper # | NLC2006-60,SP2006-116 |
| Date of Issue |
| Conference Information | |
| Committee | NLC |
|---|---|
| Conference Date | 2006/12/15(1days) |
| Place (in Japanese) | (See Japanese page) |
| Place (in English) | |
| Topics (in Japanese) | (See Japanese page) |
| Topics (in English) | |
| Chair | |
| Vice Chair | |
| Secretary | |
| Assistant | |
| Paper Information | |
| Registration To | Natural Language Understanding and Models of Communication (NLC) |
|---|---|
| Language | ENG |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | Utilizing Bayesian Network and Junction Tree Decomposition for Incorporating Additional Knowledge Sources into a Statistical Acoustic Model |
| Sub Title (in English) | |
| Keyword(1) | Acoustic modeling |
| Keyword(2) | knowledge incorporation |
| Keyword(3) | Bayesian network |
| Keyword(4) | junction tree |
| Keyword(5) | wide-context dependency |
| 1st Author's Name | Sakriani SAKTI |
| 1st Author's Affiliation | National Institute of Information and Communication Technology:ATR Spoken Language Communication Research Laboratories() |
| 2nd Author's Name | Konstantin MARKOV |
| 2nd Author's Affiliation | National Institute of Information and Communication Technology:ATR Spoken Language Communication Research Laboratories |
| 3rd Author's Name | Satoshi NAKAMURA |
| 3rd Author's Affiliation | National Institute of Information and Communication Technology:ATR Spoken Language Communication Research Laboratories |
| Date | 2006/12/15 |
| Paper # | NLC2006-60,SP2006-116 |
| Volume (vol) | vol.106 |
| Number (no) | 442 |
| Page | pp.pp.- |
| #Pages | 6 |
| Date of Issue | |