2014-09-22

Leftover questions

• Investigating LOUDS FST.
• CLG embedded decoder plus online compiler.
• DNN-GMM co-training
• NN LM

AM development

Sparse DNN

• Investigating layer-based DNN training

Noise training

• First draft of the noisy training journal paper
• Check abnormal behavior with large sigma (Yinshi, Liuchao)

Drop out & Rectification & convolutive network

• Drop out

• No performance improvement found yet.
• [1]

• Rectification

• Dropout NA problem was caused by large magnitude of weights

• Convolutive network

1. Test more configurations

Denoising & Farfield ASR

• Lasso-based de-reverberation is done with the REVERBERATION toolkit

• Start to compose the experiment section for the SL paper.

VAD

• Noise model training done. Under testing.
• Need to investigate the performance reduction in babble noise. Call Jia.

Speech rate training

• Some interesting results with the simple speech rate change algorithm was obtained on the WSJ db

[2]

• Seems ROS model is superior to the normal one with faster speech
• Need to check distribution of ROS on WSJ
• Suggest to extract speech data of different ROS, construct a new test set
• Suggest to use Tencent training data
• Suggest to remove silence when compute ROS

Scoring

• Pitch & rythmn done.
• Harmonics hold

Confidence

• Basic confidence by using lattice-based posterior + DNN posterior + ROS done
• 23% detection error achieved by balanced model

LM development

Domain specific LM

h2. domain specific count dumped h2. ngram generation is on going

h2. NUM tag LM:

• HCLG union seems better than G union, when integrating grammar + LM (25->23)
• Boost specific words like wifi if TAG model does not work for a particular word.

Word2Vector

W2V based doc classification

• Initial results variable Bayesian GMM obtained. Performance is not as good as the conventional GMM.
• Non-linear inter-language transform: English-Spanish-Czch: wv model training done, transform model on investigation

• probably over-fitting with the MLP training
• SSA-based local linear mapping still on running

• Knowledge vector started

• document obtained from wiki

• Character to word conversion

• Design the transform model

RNN LM

• Prepare WSJ database
• Trained model 10000 x 4 + 320 + 10000
• Better performance obtained (4.16-3.47)
• gigaword sampling for Chinese data

Speaker ID

• Second model done

Emotion detection

• delivered to Sinovoice

Translation

• v3.0 demo released

QA

• Framework design done