Name
Title | ||
|---|---|---|
Multilingual sequence-to-sequence speech recognition: architecture, transfer learning, and language modeling. |
Abstract | ||
|---|---|---|
Sequence-to-sequence (seq2seq) approach for low-resource ASR is a relatively new direction in speech research. The approach benefits by performing model training without using lexicon and alignments. However, this poses a new problem of requiring more data compared to conventional DNN-HMM systems. In this work, we attempt to use data from 10 BABEL languages to build a multilingual seq2seq model as a prior model, and then port them towards 4 other BABEL languages using transfer learning approach. We also explore different architectures for improving the prior multilingual seq2seq model. The paper also discusses the effect of integrating a recurrent neural network language model (RNNLM) with a seq2seq model during decoding. Experimental results show that the transfer learning approach from the multilingual model shows substantial gains over monolingual models across all 4 BABEL languages. Incorporating an RNNLM also brings significant improvements in terms of %WER, and achieves recognition performance comparable to the models trained with twice more training data. |
Year | Venue | Keywords |
|---|---|---|
2018 | 2018 IEEE Spoken Language Technology Workshop (SLT) | Decoding,Training,Data models,Mathematical model,Convolution,Two dimensional displays,Speech recognition |
DocType | Volume | ISSN |
Conference | abs/1810.03459 | 2639-5479 |
Citations | PageRank | References |
1 | 0.35 | 9 |
Authors | ||
9 | ||
Name | Order | Citations | PageRank |
|---|---|---|---|
| Jaejin Cho | 1 | 8 | 2.91 |
| Murali Karthick Baskar | 2 | 8 | 4.99 |
| ruizhi li | 3 | 51 | 12.01 |
| Matthew Wiesner | 4 | 5 | 2.85 |
| Sri Harish Reddy Mallidi | 5 | 1 | 0.35 |
| Nelson Yalta | 6 | 11 | 2.17 |
| Martin Karafiát | 7 | 154 | 12.74 |
| Shinji Watanabe | 8 | 1158 | 139.38 |
| Takaaki Hori | 9 | 408 | 45.58 |