Abstract | ||
---|---|---|
We present our work on improving the numerator graph for discriminative training using the lattice-free maximum mutual information (MMI) criterion. Specifically, we propose a scheme for creating unconstrained numerator graphs by removing time constraints from the baseline numerator graphs. This leads to much smaller graphs and therefore faster preparation of training supervisions. By testing the proposed un-constrained supervisions using factorized time-delay neural network (TDNN) models, we observe 0.5% to 2.6% relative improvement over the state-of-the-art word error rates on various large-vocabulary speech recognition databases. |
Year | DOI | Venue |
---|---|---|
2018 | 10.1109/SLT.2018.8639684 | SLT |
Keywords | Field | DocType |
Hidden Markov models,Training,Switches,Lattices,Computational modeling,Time factors,Error analysis | Graph,Pattern recognition,Computer science,Speech recognition,Time delay neural network,Artificial intelligence,Mutual information,Hidden Markov model,Artificial neural network,Discriminative model,Fraction (mathematics) | Conference |
ISSN | ISBN | Citations |
2639-5479 | 978-1-5386-4334-1 | 0 |
PageRank | References | Authors |
0.34 | 0 | 5 |
Name | Order | Citations | PageRank |
---|---|---|---|
Hossein Hadian | 1 | 11 | 3.31 |
Daniel Povey | 2 | 2442 | 231.75 |
Hossein Sameti | 3 | 229 | 41.40 |
Jan Trmal | 4 | 235 | 20.91 |
Sanjeev Khudanpur | 5 | 2155 | 202.00 |