http://cslt.org/mediawiki/index.php?action=history&feed=atom&title=2014-07-18
2014-07-18 - 版本历史
2026-04-17T06:52:41Z
本wiki的该页面的版本历史
MediaWiki 1.23.3
http://cslt.org/mediawiki/index.php?title=2014-07-18&diff=10471&oldid=prev
Zhaomy：/* Multilingual ASR */
2014-07-18T02:34:30Z
<p><span dir="auto"><span class="autocomment">Multilingual ASR</span></span></p>
<table class='diff diff-contentalign-left'>
<col class='diff-marker' />
<col class='diff-content' />
<col class='diff-marker' />
<col class='diff-content' />
<tr style='vertical-align: top;'>
<td colspan='2' style="background-color: white; color:black; text-align: center;">←上一版本</td>
<td colspan='2' style="background-color: white; color:black; text-align: center;">2014年7月18日 (五) 02:34的版本</td>
</tr><tr><td colspan="2" class="diff-lineno">第26行：</td>
<td colspan="2" class="diff-lineno">第26行：</td></tr>
<tr><td class='diff-marker'> </td><td style="background-color: #f9f9f9; color: #333333; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #e6e6e6; vertical-align: top; white-space: pre-wrap;"><div>  Tel201406.v1.0.S  |        |    -    |    -    |    -    |    -    |    -    |</div></td><td class='diff-marker'> </td><td style="background-color: #f9f9f9; color: #333333; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #e6e6e6; vertical-align: top; white-space: pre-wrap;"><div>  Tel201406.v1.0.S  |        |    -    |    -    |    -    |    -    |    -    |</div></td></tr>
<tr><td class='diff-marker'> </td><td style="background-color: #f9f9f9; color: #333333; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #e6e6e6; vertical-align: top; white-space: pre-wrap;"><div>  Tel201406.v1.1.S  |    -    |    -    |    -    |    -    |    -    |    -    |</div></td><td class='diff-marker'> </td><td style="background-color: #f9f9f9; color: #333333; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #e6e6e6; vertical-align: top; white-space: pre-wrap;"><div>  Tel201406.v1.1.S  |    -    |    -    |    -    |    -    |    -    |    -    |</div></td></tr>
<tr><td class='diff-marker'>−</td><td style="color:black; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;"><div>Tel201406.HW.v2.0.B|  20.18  |  17.49  |  23.85  |  22.81  |  22.48  |  <del class="diffchange diffchange-inline">55</del>.<del class="diffchange diffchange-inline">06 </del> |</div></td><td class='diff-marker'>+</td><td style="color:black; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;"><div>Tel201406.HW.v2.0.B|  20.18  |  17.49  |  23.85  |  22.81  |  22.48  |  <ins class="diffchange diffchange-inline">83</ins>.<ins class="diffchange diffchange-inline">28 </ins> |</div></td></tr>
<tr><td class='diff-marker'>−</td><td style="color:black; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;"><div>Tel201406.HW.v2.0.S|  19.95  |  17.74  |  23.73  |  22.36  |  22.49  |  <del class="diffchange diffchange-inline">37</del>.<del class="diffchange diffchange-inline">63 </del> |</div></td><td class='diff-marker'>+</td><td style="color:black; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;"><div>Tel201406.HW.v2.0.S|  19.95  |  17.74  |  23.73  |  22.36  |  22.49  |  <ins class="diffchange diffchange-inline">67</ins>.<ins class="diffchange diffchange-inline">70 </ins> |</div></td></tr>
<tr><td class='diff-marker'>−</td><td style="color:black; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;"><div>Tel201406.HW.v2.1.B|  19.14  |  16.97  |  24.26  |  22.28  |  22.97  |  <del class="diffchange diffchange-inline">55</del>.<del class="diffchange diffchange-inline">35 </del> |</div></td><td class='diff-marker'>+</td><td style="color:black; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;"><div>Tel201406.HW.v2.1.B|  19.14  |  16.97  |  24.26  |  22.28  |  22.97  |  <ins class="diffchange diffchange-inline">85</ins>.<ins class="diffchange diffchange-inline">41 </ins> |</div></td></tr>
<tr><td class='diff-marker'>−</td><td style="color:black; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;"><div>Tel201406.HW.v2.1.S|  19.44  |  17.62  |  24.49  |  23.06  |  23.60  |  <del class="diffchange diffchange-inline">44</del>.<del class="diffchange diffchange-inline">81 </del> |</div></td><td class='diff-marker'>+</td><td style="color:black; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;"><div>Tel201406.HW.v2.1.S|  19.44  |  17.62  |  24.49  |  23.06  |  23.60  |  <ins class="diffchange diffchange-inline">74</ins>.<ins class="diffchange diffchange-inline">61 </ins> |</div></td></tr>
<tr><td class='diff-marker'> </td><td style="background-color: #f9f9f9; color: #333333; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #e6e6e6; vertical-align: top; white-space: pre-wrap;"><div></pre></div></td><td class='diff-marker'> </td><td style="background-color: #f9f9f9; color: #333333; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #e6e6e6; vertical-align: top; white-space: pre-wrap;"><div></pre></div></td></tr>
<tr><td class='diff-marker'> </td><td style="background-color: #f9f9f9; color: #333333; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #e6e6e6; vertical-align: top; white-space: pre-wrap;"></td><td class='diff-marker'> </td><td style="background-color: #f9f9f9; color: #333333; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #e6e6e6; vertical-align: top; white-space: pre-wrap;"></td></tr>
</table>
Zhaomy
http://cslt.org/mediawiki/index.php?title=2014-07-18&diff=10469&oldid=prev
Cslt：以内容“==Resoruce Building== == Leftover questions== * Asymmetric window: Great improvement on training set(WER 34% to 24%), however the improvement is lost on test. * Multi...”创建新页面
2014-07-18T01:54:13Z
<p>以内容“==Resoruce Building== == Leftover questions== * Asymmetric window: Great improvement on training set(WER 34% to 24%), however the improvement is lost on test. * Multi...”创建新页面</p>
<p><b>新页面</b></p><div>==Resoruce Building==<br />
<br />
== Leftover questions==<br />
* Asymmetric window: Great improvement on training set(WER 34% to 24%), however the improvement is lost on test. <br />
* Multi GPU training: Error encountered<br />
* Multilanguage training<br />
* Investigating LOUDS FST. <br />
* CLG embedded decoder plus online compiler.<br />
* DNN-GMM co-training<br />
<br />
== AM development ==<br />
<br />
=== Sparse DNN ===<br />
* GA-based block sparsity (++++++++++)<br />
<br />
===Noise training===<br />
:* Journal paper writing on going<br />
<br />
===Multilingual ASR===<br />
<br />
<pre><br />
<br />
LM = Tel201406.HW.v2.1.1<br />
<br />
AM\testset |JS27H_100| JS_2h |ShanXi_2h|ShaanXi2h|Unknown2h| ENG |<br />
Tel201406.v1.0.S | | - | - | - | - | - |<br />
Tel201406.v1.1.S | - | - | - | - | - | - |<br />
Tel201406.HW.v2.0.B| 20.18 | 17.49 | 23.85 | 22.81 | 22.48 | 55.06 |<br />
Tel201406.HW.v2.0.S| 19.95 | 17.74 | 23.73 | 22.36 | 22.49 | 37.63 |<br />
Tel201406.HW.v2.1.B| 19.14 | 16.97 | 24.26 | 22.28 | 22.97 | 55.35 |<br />
Tel201406.HW.v2.1.S| 19.44 | 17.62 | 24.49 | 23.06 | 23.60 | 44.81 |<br />
</pre><br />
<br />
* v1.*: no English words involved.<br />
* v2.*: with English words involved.<br />
<br />
===Denoising & Farfield ASR===<br />
<br />
* Sparse linear prediction. Need to correct the model.<br />
* Use xEnt as the adaptation object, instead of MSE based feature mapping<br />
* Use the simulation tool to add reverberation. <br />
* [http://reverb2014.dereverberation.com/download.html]<br />
* Investigate the impact of speech rate. Use Tencent 200h data to conduct the experiments.<br />
* Investigate the correlation between phone speed & entropy.<br />
<br />
===VAD===<br />
<br />
* Waiting for engineering work<br />
<br />
===Scoring===<br />
<br />
* Refine the acoustic model with AMIDA database. problem solved by involving both wsj and AMIDA.<br />
* Model ready for picking up<br />
<br />
<br />
===Embedded decoder===<br />
<br />
* The first deliver is Emb201407_BG_v0.0<br />
* Train two smaller network: 500x4+600, 400x4+500<br />
<br />
==LM development==<br />
<br />
===Domain specific LM===<br />
<br />
h2. Domain specific LM construction<br />
<br />
h3. TAG LM<br />
* Some problems with the tagging. all numbers are tagged.<br />
<br />
h3. Chatting LM<br />
* Building chatting lexicon<br />
<br />
<br />
==Word2Vector==<br />
<br />
===W2V based doc classification===<br />
<br />
* Initial results variable Bayesian GMM obtained. Performance is not as good as the conventional GMM.<br />
* <br />
<br />
==Semantic word tree==<br />
<br />
:* Version v2.0 released (filter with query log)<br />
:* Please deliver to /nfs/disk/perm/data/corpora/semanticTree (Xingchao)<br />
:* Version v3.0 under going. Further refinement with Baidu Baike hierarchy<br />
<br />
<br />
==NN LM==<br />
<br />
* Character-based NNLM (6700 chars, 7gram), 500M data training done.<br />
:* Inconsistent pattern in WER were found on Tenent test sets<br />
:* probably need to use another test set to do investigation. <br />
<br />
* Investigate MS RNN LM training<br />
<br />
==Speaker ID==<br />
* reading materials <br />
* prepare to run sre08<br />
<br />
==Translation==<br />
* Initial version released<br />
* collecting more data (Xinhua parallel text, bible, name entity) for the second version</div>
Cslt