Name
Title | ||
|---|---|---|
Discriminative Training Based on the Criterion of Least Phone Competing Tokens for Large Vocabulary Speech Recognition |
Abstract | ||
|---|---|---|
In this paper, we propose a new discriminative training approach based on the criterion of least phone competing tokens. In our approach, we first collect a competing token set for each physical HMM from training data. Different from the previous in-search token selection, an off-line token collection procedure is used in this work to collect the competing-tokens from word lattices. Then we re-estimate HMM parameters discriminatively to minimize the total number of competing tokens counted in the phone level. The phone token counts are approximated by a sigmoid-based objective function. The GPD algorithm is used to adjust HMM parameters to minimize the objective function. In this work, a merging mechanism and a gradient normalization in the HMM tied-state level are proposed to improve the generalization power of our discriminative training method. The proposed method is evaluated on the Resource Management (RM) and the Switchboard (a 24-hr mini-train set) tasks. Experimental results clearly show that our new discriminative training method achieves significant improvements over our best MLE models in both tasks, namely about 8% and 4.5% relative error rate reduction in RM and Switchboard respectively, over the best MLE models. © 2005 IEEE. |
Year | DOI | Venue |
|---|---|---|
2005 | 10.1109/ICASSP.2005.1415064 | ICASSP (1) |
Keywords | DocType | Volume |
competing token set,off-line token collection procedure,least phone competing tokens,speech recognition,learning (artificial intelligence),relative error rate reduction,acoustic models,hmm,large vocabulary speech recognition,merging mechanism,discriminative training method,gradient normalization,gradient methods,gpd algorithm,word lattices,hidden markov models,sigmoid-based objective function minimization,merging,maximum likelihood estimation,lattices,automatic speech recognition,learning artificial intelligence | Conference | 1 |
Issue | ISSN | ISBN |
null | 1520-6149 | 0-7803-8874-7 |
Citations | PageRank | References |
2 | 0.48 | 6 |
Authors | ||
4 | ||
Name | Order | Citations | PageRank |
|---|---|---|---|
| Bo Liu | 1 | 2 | 0.48 |
| Hui Jiang | 2 | 1493 | 113.16 |
| Jian-Lai Zhou | 3 | 184 | 20.85 |
| Ren-Hua Wang | 4 | 344 | 41.36 |