Sinovoice-2016-4-21

Data

• 16K LingYun

• 2000h data ready
• 4300h real-env data to label

• YueYu

• Total 250h(190h-YueYu + 60h-English)
• Add 60h YueYu
• CER: 75%->76%

• WeiYu

• 50h for training
• 120h labeled ready

Deletion Error Promblem

• Add one noise phone to alleviate the silence over-training
• Omit sil accuracy in discriminative training

• Testdata: test_1000ju from 8000ju

  ---------------------------------------------------
                 model   | ins  |  del  | sub | wer  
  ---------------------------------------------------
         baseMPE 3.mdl   |  25  |  68   | 468 | 9.50
  ---------------------------------------------------
   MPE omit_sil_acc 3.mdl|  26  |  72   | 453 | 9.33
  ---------------------------------------------------

• Testdata: test_2000ju from 10000ju

  ----------------------------------------------------
                 model   | ins  |  del  | sub  | wer  
  ----------------------------------------------------
         baseMPE 3.mdl   |  96  |  768  | 1590 | 19.39
  ----------------------------------------------------
  MPE omit_sil_acc 3.mdl |  165 |  627  | 1685 | 19.58
  ----------------------------------------------------

• H smoothing of XEnt and MPE
• Add one silence arc from start-state to end-state

Big-Model Training

• 7*2048-10000h net weight-matrix factoring, to improve the decoding speed --SVD

• SVD looks OK, but fine-tuning still didn't work.

 Base WER:
 relu_2000_mpe_1000H: 17.72
 relu_1200_mpe_1000H: 18.60

 |layer / nodes retaind|  200  |  400  |  600  |  800  | 1000  | 1200  | 1400  |  1600  |
 |      hidden 2       |       |       | 22.53 | 20.30 | 19.01 |       |       |        |
 |      hidden 7       |       | 18.92 | 18.30 | 17.92 |       |       |       |        |     
 |       final         |       |       | 18.32 | 18.00 | 17.83 |       |       |        |     

• 7*1024 cross-entropy total train, then mpe, 0.2 improvment
• 7*1024 svd factoring, speed the decoding

• 8k

Embedding

• 10000h-chain 5*400+800 DONE.

• Beam affect the performance of chain model significantly, need more investigation.

• 5*576-2400 TDNN model

SinSong Robot

• Test based on 10000h(7*2048-xent) model

 ------------------------------------------------
   condition | clean  | replay(0.5m) | real-env
 ------------------------------------------------
     wer     |   3    |  18(mpe-14)  | too-bad
 ------------------------------------------------

• Plan to record in restaurant on April 10.

Character LM

• Except Sogou-2T, 9-gram has been done.
• Worse than word-lm(9%->6%)
• Add word boundary tag to Character-LM trainig
• Merge Character-LM & word-LM

• Union
• Compose, success.

• 2-step decoding: first, character-based LM. Then, word-based LM.

Project

• Pingan & Yueyu Deletion error too more

• TDNN deletion error rate > DNN deletion error rate
• TDNN Silence scale is too sensitive for different test cases.

SID

Digit

• Same Channel test EER: 100%

• Speaker confirm
• phone channel

• Cross Channel

• Mic-wav PLDA adaptation EER from 9% to 7% (20-30 persons)