http://index.cslt.org/mediawiki/api.php?action=feedcontributions&feedformat=atom&user=Ltycslt Wiki - 用户贡献 [zh-cn]2026-04-07T11:18:56Z用户贡献MediaWiki 1.23.3http://index.cslt.org/mediawiki/index.php/Vivi-poem-generationVivi-poem-generation2018-07-24T17:05:02Z<p>Lty：</p>
<hr />
<div>=薇薇：会写诗的机器人=<br />
<br />
成员：王东，王琪鑫，骆天一，张纪袁，冯洋<br />
<br />
==vivi 3.0 (on going) ==<br />
<br />
===目标===<br />
<br />
* Transfer modern sentences to poems<br />
* Utilize extra knowledge to boost innovation<br />
* Reinforcement learning to improve quality<br />
<br />
<br />
==vivi 2.0==<br />
<br />
===基本方法===<br />
<br />
* Tensorflow 实现<br />
* Attention-based LSTM/GRU S2S<br />
* Sampling words as input to generate the present sentence<br />
* Memory augmentation (global and local)<br />
* Local attention for theme (+)<br />
* Local attention on previous generation, with couplet assignment （line number?） (+)<br />
* N-best decoding (+)<br />
<br />
===实现细节===<br />
<br />
* Rythms with less characters removed<br />
* Characters seldom used as rhythms words are removed<br />
* Characters that are low-frequency are removed <br />
<br />
===特性===<br />
<br />
* 训练基础模型，用memory实现精细创新<br />
* 用memory可实现风格、体例转换<br />
* 用Local attention可实现人为指导创作（+）<br />
* 可实现律诗中的对仗<br />
<br />
===测试结果===<br />
<br />
===论文===<br />
<br />
* [https://arxiv.org/abs/1705.03773 Creative generation of poems]<br />
<br />
==vivi 1.0==<br />
<br />
===基本方法===<br />
<br />
* Theano 实现<br />
* 基于sequence-to-sequence的LSTM/GRU模型, 运用Attention 机制。<br />
* 输入为一首诗的第一句，输出为后面所有句子<br />
* 预训练word vectors，用多种体例古文结合在一起训练<br />
* 生成时可对用户输入进行扩展<br />
<br />
<br />
<br />
===测试结果===<br />
<br />
* [[中国古诗词图灵测试|vivi 1.0 图灵测试结果]]<br />
<br />
===论文===<br />
<br />
* [https://arxiv.org/abs/1604.06274|Chinese Song Iambics Generation with Neural Attention-based Model, IJCAI2016]<br />
<br />
* [http://link.springer.com/chapter/10.1007/978-3-319-49685-6_4/fulltext.html|Can Machine Generate Traditional Chinese Poetry? A Feigenbaum Test, Springer, LNCS, vol 10023, pp.171-183.]<br />
<br />
* [https://arxiv.org/abs/1705.03773 Jiyuan Zhang, Yang Feng, Dong Wang, Yang Wang, Andrew Abel, Shiyue Zhang, Andi Zhangi, "Flexible and Creative Chinese Poetry Generation Using Neural Memory"]<br />
<br />
<br />
===文章===<br />
<br />
[[Wangd-wiki-article-vvpoem|薇薇的故事]]</div>Ltyhttp://index.cslt.org/mediawiki/index.php/Tianyi_LuoTianyi Luo2017-08-15T02:38:43Z<p>Lty：</p>
<hr />
<div>'''Tainyi Luo(骆天一)'''<br />
<br />
'''Education:'''<br />
<br />
Master: Peking University(北京大学) (2013).<br />
<br />
'''Work experience：'''<br />
* 2014.12-: Machine Learning/Natural Language Processing Research Engineer in CSLT, RIIT, Tsinghua Univ., China. (清华大学) (Advisor: [http://wangd.cslt.org/ Dong Wang]])<br />
<br />
'''Research interests:''' <br />
<br />
Machine Learning, Natural Language Processing, Information Retrieval and Recommend System<br />
<br />
'''Publications([http://lty.cslt.org/ Tianyi Luo's Homepage]):''' <br />
#Tianyi Luo*, Qixin Wang*, Dong Wang. “Chinese Song Iambics Generation with Neural Attention-based Model”, The 30th AAAI Conference on Artificial Intelligence (AAAI 2016), full paper submitted. (*: equal contribution) [[http://cslt.riit.tsinghua.edu.cn/mediawiki/images/3/34/Songci.pdf Paper]]<br />
#Tianyi Luo, Dong Wang, Rong Liu and Yiqiao Pan, "Stochastic Top-k ListNet", Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing (EMNLP 2015 long oral paper) , pp. 676-684, Sep 17-21, 2015. Lisbon, Portugal. [[http://www.aclweb.org/anthology/D/D15/D15-1079.pdf Paper]] [[http://cslt.riit.tsinghua.edu.cn/mediawiki/images/0/00/Emnlp2015longslides.pdf PPT]] [[https://github.com/pkuluotianyi/topKStoListNet Code]]<br />
#Dongxu Zhang, Tianyi Luo, Rong Liu, Dong Wang. “Learning from LDA using Deep Neural Networks”, arXiv: 1508.01011 [[http://arxiv.org/pdf/1508.01011.pdf Paper]] [[http://cslt.riit.tsinghua.edu.cn/mediawiki/images/e/ec/Emnlp2015shortslides.pdf PPT]] [[http://pan.baidu.com/s/1i3Ek35b Code]]<br />
<br />
'''Related documents'''<br />
*Tianyi Luo's Moses installing and training process [[http://cslt.riit.tsinghua.edu.cn/mediawiki/images/f/fe/2016-01-03_Moses%E5%AE%89%E8%A3%85%E8%AE%AD%E7%BB%83%E5%85%A8%E8%BF%87%E7%A8%8B.pdf pdf]] 01/04/2016<br />
*Tianyi Luo's yearly report(01/2015~12/2015) [[http://cslt.riit.tsinghua.edu.cn/mediawiki/images/4/4c/2016-01-06_Tianyi_Luo%27s_yearly_report.pdf pdf]] 01/06/2016<br />
*Intellegent QA System Overview [[http://cslt.riit.tsinghua.edu.cn/mediawiki/images/7/7d/Cslt-trp-template.pdf pdf]] 03/06/2016</div>Ltyhttp://index.cslt.org/mediawiki/index.php/Tianyi_LuoTianyi Luo2017-08-15T02:38:12Z<p>Lty：</p>
<hr />
<div>'''Tainyi Luo(骆天一)'''<br />
<br />
'''Education:'''<br />
<br />
Master: Peking University(北京大学) (2013).<br />
<br />
'''Work experience：'''<br />
* 2014.5-2014.9: Research Assistant in Natural Language Processing, The Hong Kong Polytechnic University, Hong Kong, China (香港理工大学) (Advisor: [http://www4.comp.polyu.edu.hk/~cswjli/ Wenjie Li]])<br />
* 2014.12-: Machine Learning/Natural Language Processing Research Engineer in CSLT, RIIT, Tsinghua Univ., China. (清华大学) (Advisor: [http://wangd.cslt.org/ Dong Wang]])<br />
<br />
'''Research interests:''' <br />
<br />
Machine Learning, Natural Language Processing, Information Retrieval and Recommend System<br />
<br />
'''Publications([http://lty.cslt.org/ Tianyi Luo's Homepage]):''' <br />
#Tianyi Luo*, Qixin Wang*, Dong Wang. “Chinese Song Iambics Generation with Neural Attention-based Model”, The 30th AAAI Conference on Artificial Intelligence (AAAI 2016), full paper submitted. (*: equal contribution) [[http://cslt.riit.tsinghua.edu.cn/mediawiki/images/3/34/Songci.pdf Paper]]<br />
#Tianyi Luo, Dong Wang, Rong Liu and Yiqiao Pan, "Stochastic Top-k ListNet", Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing (EMNLP 2015 long oral paper) , pp. 676-684, Sep 17-21, 2015. Lisbon, Portugal. [[http://www.aclweb.org/anthology/D/D15/D15-1079.pdf Paper]] [[http://cslt.riit.tsinghua.edu.cn/mediawiki/images/0/00/Emnlp2015longslides.pdf PPT]] [[https://github.com/pkuluotianyi/topKStoListNet Code]]<br />
#Dongxu Zhang, Tianyi Luo, Rong Liu, Dong Wang. “Learning from LDA using Deep Neural Networks”, arXiv: 1508.01011 [[http://arxiv.org/pdf/1508.01011.pdf Paper]] [[http://cslt.riit.tsinghua.edu.cn/mediawiki/images/e/ec/Emnlp2015shortslides.pdf PPT]] [[http://pan.baidu.com/s/1i3Ek35b Code]]<br />
<br />
'''Related documents'''<br />
*Tianyi Luo's Moses installing and training process [[http://cslt.riit.tsinghua.edu.cn/mediawiki/images/f/fe/2016-01-03_Moses%E5%AE%89%E8%A3%85%E8%AE%AD%E7%BB%83%E5%85%A8%E8%BF%87%E7%A8%8B.pdf pdf]] 01/04/2016<br />
*Tianyi Luo's yearly report(01/2015~12/2015) [[http://cslt.riit.tsinghua.edu.cn/mediawiki/images/4/4c/2016-01-06_Tianyi_Luo%27s_yearly_report.pdf pdf]] 01/06/2016<br />
*Intellegent QA System Overview [[http://cslt.riit.tsinghua.edu.cn/mediawiki/images/7/7d/Cslt-trp-template.pdf pdf]] 03/06/2016</div>Ltyhttp://index.cslt.org/mediawiki/index.php/News-20161008News-201610082016-10-08T06:08:32Z<p>Lty：</p>
<hr />
<div>==Members left CSLT recently==<br />
<br />
===Caixia Wang===<br />
<br />
Special Visitor, left for University of Diplomacy <br />
<br />
===Maoning Wang===<br />
<br />
Special Visitor, left for University of Central Economics and Finance <br />
<br />
===[http://lty.cslt.org/ Tianyi Luo]=== <br />
<br />
Research engineer, left for University of California, Santa Cruz<br />
<br />
[[文件:10.pic.jpg|200px]]<br />
<br />
===Xiangyu Zeng=== <br />
<br />
Visiting students, left for University of Columbia <br />
<br />
===Yiqiao Pan=== <br />
<br />
Visiting student, left for University of Sydney<br />
<br />
===Qixin Wang=== <br />
<br />
Visiting student, left for University of South California<br />
<br />
==New members joined CSLT recently==<br />
<br />
===Yang Feng=== <br />
Special visitor<br />
<br />
NLP<br />
<br />
===Yang Wang=== <br />
Student<br />
<br />
Financial processing<br />
<br />
===Xingliang Cheng=== <br />
<br />
Student<br />
<br />
Speech processing<br />
<br />
===Shiyue Zhang===<br />
Visitor<br />
<br />
NLP<br />
<br />
<br />
===Jiyuan Zhang===<br />
Visitor<br />
<br />
NLP<br />
<br />
<br />
===Ying Shi===<br />
Visitor<br />
<br />
Speech processing<br />
<br />
===Shiyao LI===<br />
Intern<br />
<br />
NLP<br />
<br />
===Yixang Chen===<br />
Intern<br />
<br />
Speech processing<br />
<br />
<br />
===Aodong Li===<br />
Visitor<br />
<br />
NLP<br />
<br />
<br />
===Jingyi Lin===<br />
<br />
Visitor<br />
<br />
Speech processing</div>Ltyhttp://index.cslt.org/mediawiki/index.php/News-20161008News-201610082016-10-08T06:04:29Z<p>Lty：</p>
<hr />
<div>==Members left CSLT recently==<br />
<br />
===Caixia Wang===<br />
<br />
Special Visitor, left for University of Diplomacy <br />
<br />
===Maoning Wang===<br />
<br />
Special Visitor, left for University of Central Economics and Finance <br />
<br />
===Tianyi Luo=== <br />
<br />
Research engineer, left for University of California, Santa Cruz<br />
<br />
[[文件:10.pic.jpg|100px]]<br />
<br />
===Xiangyu Zeng=== <br />
<br />
Visiting students, left for University of Columbia <br />
<br />
===Yiqiao Pan=== <br />
<br />
Visiting student, left for University of Sydney<br />
<br />
===Qixin Wang=== <br />
<br />
Visiting student, left for University of South California<br />
<br />
==New members joined CSLT recently==<br />
<br />
===Yang Feng=== <br />
Special visitor<br />
<br />
NLP<br />
<br />
===Yang Wang=== <br />
Student<br />
<br />
Financial processing<br />
<br />
===Xingliang Cheng=== <br />
<br />
Student<br />
<br />
Speech processing<br />
<br />
===Shiyue Zhang===<br />
Visitor<br />
<br />
NLP<br />
<br />
<br />
===Jiyuan Zhang===<br />
Visitor<br />
<br />
NLP<br />
<br />
<br />
===Ying Shi===<br />
Visitor<br />
<br />
Speech processing<br />
<br />
===Shiyao LI===<br />
Intern<br />
<br />
NLP<br />
<br />
===Yixang Chen===<br />
Intern<br />
<br />
Speech processing<br />
<br />
<br />
===Aodong Li===<br />
Visitor<br />
<br />
NLP<br />
<br />
<br />
===Jingyi Lin===<br />
<br />
Visitor<br />
<br />
Speech processing</div>Ltyhttp://index.cslt.org/mediawiki/index.php/News-20161008News-201610082016-10-08T06:03:59Z<p>Lty：</p>
<hr />
<div>==Members left CSLT recently==<br />
<br />
===Caixia Wang===<br />
<br />
Special Visitor, left for University of Diplomacy <br />
<br />
===Maoning Wang===<br />
<br />
Special Visitor, left for University of Central Economics and Finance <br />
<br />
===Tianyi Luo=== <br />
<br />
Research engineer, left for University of California, Santa Cruz<br />
<br />
[[文件:http://cslt.riit.tsinghua.edu.cn/mediawiki/images/6/68/10.pic.jpg]]<br />
[[文件:10.pic.jpg|200px]]<br />
<br />
===Xiangyu Zeng=== <br />
<br />
Visiting students, left for University of Columbia <br />
<br />
===Yiqiao Pan=== <br />
<br />
Visiting student, left for University of Sydney<br />
<br />
===Qixin Wang=== <br />
<br />
Visiting student, left for University of South California<br />
<br />
==New members joined CSLT recently==<br />
<br />
===Yang Feng=== <br />
Special visitor<br />
<br />
NLP<br />
<br />
===Yang Wang=== <br />
Student<br />
<br />
Financial processing<br />
<br />
===Xingliang Cheng=== <br />
<br />
Student<br />
<br />
Speech processing<br />
<br />
===Shiyue Zhang===<br />
Visitor<br />
<br />
NLP<br />
<br />
<br />
===Jiyuan Zhang===<br />
Visitor<br />
<br />
NLP<br />
<br />
<br />
===Ying Shi===<br />
Visitor<br />
<br />
Speech processing<br />
<br />
===Shiyao LI===<br />
Intern<br />
<br />
NLP<br />
<br />
===Yixang Chen===<br />
Intern<br />
<br />
Speech processing<br />
<br />
<br />
===Aodong Li===<br />
Visitor<br />
<br />
NLP<br />
<br />
<br />
===Jingyi Lin===<br />
<br />
Visitor<br />
<br />
Speech processing</div>Ltyhttp://index.cslt.org/mediawiki/index.php/News-20161008News-201610082016-10-08T06:01:09Z<p>Lty：</p>
<hr />
<div>==Members left CSLT recently==<br />
<br />
===Caixia Wang===<br />
<br />
Special Visitor, left for University of Diplomacy <br />
<br />
===Maoning Wang===<br />
<br />
Special Visitor, left for University of Central Economics and Finance <br />
<br />
===Tianyi Luo=== <br />
<br />
Research engineer, left for University of California, Santa Cruz<br />
<br />
[[文件:http://cslt.riit.tsinghua.edu.cn/mediawiki/images/6/68/10.pic.jpg]]<br />
<br />
===Xiangyu Zeng=== <br />
<br />
Visiting students, left for University of Columbia <br />
<br />
===Yiqiao Pan=== <br />
<br />
Visiting student, left for University of Sydney<br />
<br />
===Qixin Wang=== <br />
<br />
Visiting student, left for University of South California<br />
<br />
==New members joined CSLT recently==<br />
<br />
===Yang Feng=== <br />
Special visitor<br />
<br />
NLP<br />
<br />
===Yang Wang=== <br />
Student<br />
<br />
Financial processing<br />
<br />
===Xingliang Cheng=== <br />
<br />
Student<br />
<br />
Speech processing<br />
<br />
===Shiyue Zhang===<br />
Visitor<br />
<br />
NLP<br />
<br />
<br />
===Jiyuan Zhang===<br />
Visitor<br />
<br />
NLP<br />
<br />
<br />
===Ying Shi===<br />
Visitor<br />
<br />
Speech processing<br />
<br />
===Shiyao LI===<br />
Intern<br />
<br />
NLP<br />
<br />
===Yixang Chen===<br />
Intern<br />
<br />
Speech processing<br />
<br />
<br />
===Aodong Li===<br />
Visitor<br />
<br />
NLP<br />
<br />
<br />
===Jingyi Lin===<br />
<br />
Visitor<br />
<br />
Speech processing</div>Ltyhttp://index.cslt.org/mediawiki/index.php/%E6%96%87%E4%BB%B6:10.pic.jpg文件:10.pic.jpg2016-10-08T06:00:21Z<p>Lty：</p>
<hr />
<div></div>Ltyhttp://index.cslt.org/mediawiki/index.php/%E6%96%87%E4%BB%B6:9.pic.jpg文件:9.pic.jpg2016-10-08T05:58:11Z<p>Lty：</p>
<hr />
<div></div>Ltyhttp://index.cslt.org/mediawiki/index.php/News-20161008News-201610082016-10-08T05:55:05Z<p>Lty：/* Members left CSLT recently */</p>
<hr />
<div>==Members left CSLT recently==<br />
<br />
===Caixia Wang===<br />
<br />
Special Visitor, left for University of Diplomacy <br />
<br />
===Maoning Wang===<br />
<br />
Special Visitor, left for University of Central Economics and Finance <br />
<br />
===Tianyi Luo=== <br />
<br />
Research engineer, left for University of California, Santa Cruz<br />
<br />
===Xiangyu Zeng=== <br />
<br />
Visiting students, left for University of Columbia <br />
<br />
===Yiqiao Pan=== <br />
<br />
Visiting student, left for University of Sydney<br />
<br />
===Qixin Wang=== <br />
<br />
Visiting student, left for University of South California<br />
<br />
==New members joined CSLT recently==<br />
<br />
===Yang Feng=== <br />
Special visitor<br />
<br />
NLP<br />
<br />
===Yang Wang=== <br />
Student<br />
<br />
Financial processing<br />
<br />
===Xingliang Cheng=== <br />
<br />
Student<br />
<br />
Speech processing<br />
<br />
===Shiyue Zhang===<br />
Visitor<br />
<br />
NLP<br />
<br />
<br />
===Jiyuan Zhang===<br />
Visitor<br />
<br />
NLP<br />
<br />
<br />
===Ying Shi===<br />
Visitor<br />
<br />
Speech processing<br />
<br />
===Shiyao LI===<br />
Intern<br />
<br />
NLP<br />
<br />
===Yixang Chen===<br />
Intern<br />
<br />
Speech processing<br />
<br />
<br />
===Aodong Li===<br />
Visitor<br />
<br />
NLP<br />
<br />
<br />
===Jingyi Lin===<br />
<br />
Visitor<br />
<br />
Speech processing</div>Ltyhttp://index.cslt.org/mediawiki/index.php/Publication-trpPublication-trp2016-06-20T19:37:32Z<p>Lty：</p>
<hr />
<div>[[文件:Aikefu.bmp|200px]]<br />
*[[媒体文件:Cslt-trp-template.pdf|TRP-20160004: A Review of Neural QA, Tianyi Luo and Dong Wang]]<br />
<br />
[[文件:simpair.png|200px]]<br />
*[[媒体文件:Trp20160003.pdf|TRP-20160003: A study of Similar Word Model for Unfrequent Word Enhancement in Speech Recognition, Xi Ma, Dong Wang and Javier Tejedor]]<br />
<br />
[[文件:low-freq.png|200px]]<br />
*[[媒体文件:How to deal with low frequency words.pdf|TRP-20160002: Low-Frequency Words Embedding, Chao Xing, Yiqiao Pan, Dong Wang]]<br />
[[文件:maxmargin.png|200px]]<br />
*[[媒体文件:Max-margin.pdf|TRP-20160001: Max-margin metric learning for speaker recognition, Lantian Li, Chao Xing, Dong Wang]]<br />
<br />
<br />
[[文件:lowv.png|200px]]<br />
*[[媒体文件:Lowv.pdf|TRP-20150033: Learning Ordered Word Representations, Xiaoxi Wang, Chao Xing, Dong Wang, Rong Liu and Yiqiao Pan]]<br />
<br />
[[文件:Adamax.png|200px]]<br />
*[[媒体文件:Adamax Online Training for Speech Recognition.pdf|TRP-20150032: Adamax Online Training for Speech Recognition, Xiangyu Zeng, Zhiyong Zhang, Dong Wang]]<br />
<br />
[[文件:Ptrnets.png|200px]]<br />
*[[媒体文件: Ptrnets.pdf|TRP-20150031: An implementation of Pointer-Networks with Extensions, Xiaoxi Wang, Dong Wang]]<br />
<br />
[[文件:dvad.png|200px]]<br />
*[[媒体文件:dvad.pdf|TRP-20150030: DNN-based Voice Activity Detection for Speaker Recognition, Fanhu Bie, Zhiyong Zhang, Dong Wang, Thomas Fang Zheng]]<br />
<br />
[[文件:uyghur.jpg|200px]]<br />
*[[媒体文件:urghur.pdf|TRP-20150029: THUYG-20：A Free Uyghur Speech Database, Askar Rozi, Shi Yin, Zhiyong Zhang, Dong Wang, Askar Hamdulla]]<br />
<br />
[[文件:nnpre.jpg|200px]]<br />
*[[媒体文件:nnpre.pdf|TRP-20150028: Knowledge Transfer Pre-training, Zhiyuang Tang, Dong Wang]]<br />
<br />
[[文件:mmargin.png|200px]]<br />
*[[媒体文件:mmargin.pdf|TRP-20150027: Max-Margin Metric Learning for Speaker Recognition, Lantian Li, Chao Xing, Dong Wang, Thomas Fang Zheng]]<br />
<br />
[[文件:binary.jpg|200px]]<br />
*[[媒体文件:binary.pdf|TRP-20150026: Binary Speaker Embedding, Lantian Li, Chao Xing, Dong Wang, Kaimin Yu, Thomas Fang Zheng]]<br />
<br />
[[文件:rnnrl.png|200px]]<br />
*[[媒体文件:rnnrl.pdf|TRP-20150025: Relation Classification via Recurrent Neural Network, Dong Xu Zhang, Dong Wang, Rong Liu]]<br />
<br />
<br />
[[文件:dplda.png|200px]]<br />
*[[媒体文件:dplda.pdf|TRP-20150024: Learning from LDA using Deep Neural Networks, Dongxu Zhang, Tianyi Luo and Dong Wang]]<br />
<br />
<br />
[[文件:jsrl.png|200px]]<br />
*[[媒体文件:jsrl.pdf|TRP-20150023: Joint Semantic Relevance Learning with Text Data and Graph Knowledge, Dongxu Zhang, Bin Yuan, Dong Wang, Rong Liu]]<br />
<br />
<br />
[[文件:listnet.png|200px]]<br />
*[[媒体文件:listnet.pdf|TRP-20150022: Stochastic Top-k ListNet, Tianyi Luo, Dong Wang, Rong Liu, Yiqiao Pan]]<br />
<br />
<br />
[[文件:segvector.png|200px]]<br />
*[[媒体文件:segvector.pdf|TRP-20150021: Improved Deep Speaker Feature Learning for Text-Dependent Speaker Recognition, Lantian Li, Yiye Lin, Zhiyong Zhang, Dong Wang]]<br />
<br />
<br />
[[文件:Vmclass.png|200px]]<br />
*[[媒体文件:Vmclass.pdf|TRP-20150020: Document Classification with Spherical Word Vectors, Yiqiao Pan, Chao Xing, Dong Wang]]<br />
<br />
<br />
[[文件:Tlearn.png|200px]]<br />
*[[媒体文件:Tlearn.pdf|TRP-20150019: Transfer Learning for Speech and Language Processing, Dong Wang and Thomas Fang Zheng]]<br />
<br />
[[文件:Songcisample.png|200px]]<br />
*[http://cslt.riit.tsinghua.edu.cn/mediawiki/images/7/7a/Cslt20150018_revisedversion.pdf TRP-20150018:Chinese Song Iambics Generation with Neural Attention-based Model, Qixin Wang, Tianyi Luo, Dong Wang, Chao Xing]<br />
<br />
[[文件:database.jpg|200px]]<br />
*[[媒体文件:Thuyg20-sre.pdf|TRP-20150017: AN OPEN/FREE DATABASE AND BENCHMARK FOR UYGHUR SPEAKER RECOGNITION, Askar Rozi, Dong Wang, Zhiyong Zhang, Thomas Fang Zheng]]<br />
<br />
[[文件:Thchs.png|200px]]<br />
*[[媒体文件:Thchs30.pdf|TRP-20150016: THCHS-30 : A Free Chinese Speech Corpus, Dong Wang and Xuewei Zhang]]<br />
<br />
<br />
[[文件:Su.jpg|200px]]<br />
*[[媒体文件:SUSR.pdf|TRP-20150015: Improving Short Utterance Speaker Recognition by Modeling Speech Unit Classes, Chenhao Zhang Dong Wang, Lantian Li and Thomas Fang Zheng]]<br />
<br />
[[文件:dv.png|200px]]<br />
*[[媒体文件:Dvector.pdf|TRP-20150014: Deep Speaker Vectors for Semi Text-independent Speaker Verification, Lantian Li, Dong Wang, Zhiyong Zhang and Thomas Fang Zheng]]<br />
<br />
[[文件:dark.png|200px]]<br />
*[[媒体文件:Dark.pdf|TRP-20150013: Recurrent Neural Network Training with Dark Knowledge Transfer, Dong Wang, Chao Liu, Zhiyuan Tang, Zhiyong Zhang, Mengyuan Zhao]]<br />
<br />
[[文件:PBE.png|200px]]<br />
*[[媒体文件:Probabilistic_Belief_Embedding_for_Knowledge_Population_(TRP).pdf|TRP-20150012: Probabilistic Belief Embedding for Large-scale Knowledge Population. Miao Fan, Qiang Zhou, Andrew Abel, Thomas Fang Zheng and Ralph Grishman]]<br />
<br />
[[文件:fst-fw.png|200px]]<br />
*[[媒体文件:wpair.pdf|TRP-20150011: Recognize Foreign Low-Frequency Words with Similar Pairs, Xi Ma1, Xiaoxi Wang1 and Dong Wang]]<br />
<br />
[[文件:Cdae.png|200px]]<br />
*[[媒体文件:Music.pdf|TRP-20150010: Music Removal by Denoising Autoencoder in Speech Recognition. Mengyuan Zhao, Dong Wang, Zhiyong Zhang and Xuewei Zhang]]<br />
<br />
[[文件:vmfsne.png|200px]]<br />
*[[媒体文件:Cslt-trp-template-vmfsne.pdf|TRP-20150009: VMF-SNE: Embedding for Spherical Data. Mian Wang, Dong Wang]]<br />
<br />
[[文件:ros.png|200px]]<br />
*[[媒体文件:Ros.pdf|TRP-20150008: Learning Speech Rate in Speech Recognition. Xiangyu Zeng, Shi Yin.]]<br />
<br />
[[文件:Dnnvadstru.png|200px]]<br />
*[[媒体文件:DNNVADTRP.pdf|TRP-20150007: Voice Activity Detection Based on Deep Neural Networks. Shi Yin.]] ([[媒体文件:Vad.pdf|Paper submiited to Tsinghua Xuebao]])<br />
<br />
[[文件:Uyghur-training.png|200px]]<br />
*[[媒体文件:UyghurTRP.pdf|TRP-20150006: Low-resource Uyghur Acoustic Model Training based on Cross-lingual Features. Shi Yin.]]<br />
<br />
[[文件:Beam-forming.png|200px]]<br />
*[[媒体文件:Multi-Microphones_Reverberation_Cancellation_for_Distant_Speech_Recognition.pdf|TRP-20150005: Multi-Microphones_Reverberation_Cancellation_for_Distant_Speech_Recognition. Xuewei Zhang.]]<br />
<br />
[[文件:Clipping-speaker.png|200px]]<br />
*[[媒体文件:Clip.pdf|TRP-20150004: Detection and Reconstruction of Clipped Speech in Speaker Recognition. Fanhu Bie et al.]]<br />
<br />
[[文件:Semi-dynamic-embedding.png|200px]]<br />
*[[媒体文件:Taglm.pdf|TRP-20150003: Semi-Dynamic Graph Embedding for Large Scale Language Model Adaptation. Bin Yuan et al.]]<br />
<br />
[[文件:Speaker-discriminative-score.png|200px]]<br />
*[[媒体文件:DNN-based Discriminative Scoring for Speaker.pdf|TRP-20150002: DNN-based Discriminative Scoring for Speaker Recognition Based on i-vector. Jun Wang et al. ]]<br />
<br />
[[文件:Noisy-traiing.png|200px]]<br />
*[[媒体文件:Noisy Training for Deep Neural Networks in.pdf|TRP-20150001: Noisy Training for Deep Neural Networks in Speech Recognition. Shi Yin et al.]]<br />
<br />
[[文件:English-scroing.png|200px]]<br />
*[[媒体文件:AutomaticScoringforEnglishUtterances.pdf|TRP-20140001: Automatic Scoring for English Utterances. Bo Hu.]]<br />
<br />
<br />
[[媒体文件:Template.rar|CSLT-TRP latex template]]</div>Ltyhttp://index.cslt.org/mediawiki/index.php/Publication-trpPublication-trp2016-06-20T19:36:03Z<p>Lty：</p>
<hr />
<div>[[文件:Aikefu.bmp|200px]]<br />
*[[媒体文件:Cslt-trp-template.pdf|TRP-20160004: A Review of Neural QA, Tianyi Luo and Dong Wang]]<br />
<br />
[[文件:simpair.png|200px]]<br />
*[[媒体文件:Trp20160003.pdf|TRP-20160003: A study of Similar Word Model for Unfrequent Word Enhancement in Speech Recognition, Xi Ma, Dong Wang and Javier Tejedor]]<br />
<br />
[[文件:low-freq.png|200px]]<br />
*[[媒体文件:How to deal with low frequency words.pdf|TRP-20160002: Low-Frequency Words Embedding, Chao Xing, Yiqiao Pan, Dong Wang]]<br />
[[文件:maxmargin.png|200px]]<br />
*[[媒体文件:Max-margin.pdf|TRP-20160001: Max-margin metric learning for speaker recognition, Lantian Li, Chao Xing, Dong Wang]]<br />
<br />
<br />
[[文件:lowv.png|200px]]<br />
*[[媒体文件:Lowv.pdf|TRP-20150033: Learning Ordered Word Representations, Xiaoxi Wang, Chao Xing, Dong Wang, Rong Liu and Yiqiao Pan]]<br />
<br />
[[文件:Adamax.png|200px]]<br />
*[[媒体文件:Adamax Online Training for Speech Recognition.pdf|TRP-20150032: Adamax Online Training for Speech Recognition, Xiangyu Zeng, Zhiyong Zhang, Dong Wang]]<br />
<br />
[[文件:Ptrnets.png|200px]]<br />
*[[媒体文件: Ptrnets.pdf|TRP-20150031: An implementation of Pointer-Networks with Extensions, Xiaoxi Wang, Dong Wang]]<br />
<br />
[[文件:dvad.png|200px]]<br />
*[[媒体文件:dvad.pdf|TRP-20150030: DNN-based Voice Activity Detection for Speaker Recognition, Fanhu Bie, Zhiyong Zhang, Dong Wang, Thomas Fang Zheng]]<br />
<br />
[[文件:uyghur.jpg|200px]]<br />
*[[媒体文件:urghur.pdf|TRP-20150029: THUYG-20：A Free Uyghur Speech Database, Askar Rozi, Shi Yin, Zhiyong Zhang, Dong Wang, Askar Hamdulla]]<br />
<br />
[[文件:nnpre.jpg|200px]]<br />
*[[媒体文件:nnpre.pdf|TRP-20150028: Knowledge Transfer Pre-training, Zhiyuang Tang, Dong Wang]]<br />
<br />
[[文件:mmargin.png|200px]]<br />
*[[媒体文件:mmargin.pdf|TRP-20150027: Max-Margin Metric Learning for Speaker Recognition, Lantian Li, Chao Xing, Dong Wang, Thomas Fang Zheng]]<br />
<br />
[[文件:binary.jpg|200px]]<br />
*[[媒体文件:binary.pdf|TRP-20150026: Binary Speaker Embedding, Lantian Li, Chao Xing, Dong Wang, Kaimin Yu, Thomas Fang Zheng]]<br />
<br />
[[文件:rnnrl.png|200px]]<br />
*[[媒体文件:rnnrl.pdf|TRP-20150025: Relation Classification via Recurrent Neural Network, Dong Xu Zhang, Dong Wang, Rong Liu]]<br />
<br />
<br />
[[文件:dplda.png|200px]]<br />
*[[媒体文件:dplda.pdf|TRP-20150024: Learning from LDA using Deep Neural Networks, Dongxu Zhang, Tianyi Luo and Dong Wang]]<br />
<br />
<br />
[[文件:jsrl.png|200px]]<br />
*[[媒体文件:jsrl.pdf|TRP-20150023: Joint Semantic Relevance Learning with Text Data and Graph Knowledge, Dongxu Zhang, Bin Yuan, Dong Wang, Rong Liu]]<br />
<br />
<br />
[[文件:listnet.png|200px]]<br />
*[[媒体文件:listnet.pdf|TRP-20150022: Stochastic Top-k ListNet, Tianyi Luo, Dong Wang, Rong Liu, Yiqiao Pan]]<br />
<br />
<br />
[[文件:segvector.png|200px]]<br />
*[[媒体文件:segvector.pdf|TRP-20150021: Improved Deep Speaker Feature Learning for Text-Dependent Speaker Recognition, Lantian Li, Yiye Lin, Zhiyong Zhang, Dong Wang]]<br />
<br />
<br />
[[文件:Vmclass.png|200px]]<br />
*[[媒体文件:Vmclass.pdf|TRP-20150020: Document Classification with Spherical Word Vectors, Yiqiao Pan, Chao Xing, Dong Wang]]<br />
<br />
<br />
[[文件:Tlearn.png|200px]]<br />
*[[媒体文件:Tlearn.pdf|TRP-20150019: Transfer Learning for Speech and Language Processing, Dong Wang and Thomas Fang Zheng]]<br />
<br />
[[文件:Songcisample.png|200px]]<br />
*[[http://cslt.riit.tsinghua.edu.cn/mediawiki/images/7/7a/Cslt20150018_revisedversion.pdf TRP-20150018:Chinese Song Iambics Generation with Neural Attention-based Model, Qixin Wang, Tianyi Luo, Dong Wang, Chao Xing]]<br />
<br />
[[文件:database.jpg|200px]]<br />
*[[媒体文件:Thuyg20-sre.pdf|TRP-20150017: AN OPEN/FREE DATABASE AND BENCHMARK FOR UYGHUR SPEAKER RECOGNITION, Askar Rozi, Dong Wang, Zhiyong Zhang, Thomas Fang Zheng]]<br />
<br />
[[文件:Thchs.png|200px]]<br />
*[[媒体文件:Thchs30.pdf|TRP-20150016: THCHS-30 : A Free Chinese Speech Corpus, Dong Wang and Xuewei Zhang]]<br />
<br />
<br />
[[文件:Su.jpg|200px]]<br />
*[[媒体文件:SUSR.pdf|TRP-20150015: Improving Short Utterance Speaker Recognition by Modeling Speech Unit Classes, Chenhao Zhang Dong Wang, Lantian Li and Thomas Fang Zheng]]<br />
<br />
[[文件:dv.png|200px]]<br />
*[[媒体文件:Dvector.pdf|TRP-20150014: Deep Speaker Vectors for Semi Text-independent Speaker Verification, Lantian Li, Dong Wang, Zhiyong Zhang and Thomas Fang Zheng]]<br />
<br />
[[文件:dark.png|200px]]<br />
*[[媒体文件:Dark.pdf|TRP-20150013: Recurrent Neural Network Training with Dark Knowledge Transfer, Dong Wang, Chao Liu, Zhiyuan Tang, Zhiyong Zhang, Mengyuan Zhao]]<br />
<br />
[[文件:PBE.png|200px]]<br />
*[[媒体文件:Probabilistic_Belief_Embedding_for_Knowledge_Population_(TRP).pdf|TRP-20150012: Probabilistic Belief Embedding for Large-scale Knowledge Population. Miao Fan, Qiang Zhou, Andrew Abel, Thomas Fang Zheng and Ralph Grishman]]<br />
<br />
[[文件:fst-fw.png|200px]]<br />
*[[媒体文件:wpair.pdf|TRP-20150011: Recognize Foreign Low-Frequency Words with Similar Pairs, Xi Ma1, Xiaoxi Wang1 and Dong Wang]]<br />
<br />
[[文件:Cdae.png|200px]]<br />
*[[媒体文件:Music.pdf|TRP-20150010: Music Removal by Denoising Autoencoder in Speech Recognition. Mengyuan Zhao, Dong Wang, Zhiyong Zhang and Xuewei Zhang]]<br />
<br />
[[文件:vmfsne.png|200px]]<br />
*[[媒体文件:Cslt-trp-template-vmfsne.pdf|TRP-20150009: VMF-SNE: Embedding for Spherical Data. Mian Wang, Dong Wang]]<br />
<br />
[[文件:ros.png|200px]]<br />
*[[媒体文件:Ros.pdf|TRP-20150008: Learning Speech Rate in Speech Recognition. Xiangyu Zeng, Shi Yin.]]<br />
<br />
[[文件:Dnnvadstru.png|200px]]<br />
*[[媒体文件:DNNVADTRP.pdf|TRP-20150007: Voice Activity Detection Based on Deep Neural Networks. Shi Yin.]] ([[媒体文件:Vad.pdf|Paper submiited to Tsinghua Xuebao]])<br />
<br />
[[文件:Uyghur-training.png|200px]]<br />
*[[媒体文件:UyghurTRP.pdf|TRP-20150006: Low-resource Uyghur Acoustic Model Training based on Cross-lingual Features. Shi Yin.]]<br />
<br />
[[文件:Beam-forming.png|200px]]<br />
*[[媒体文件:Multi-Microphones_Reverberation_Cancellation_for_Distant_Speech_Recognition.pdf|TRP-20150005: Multi-Microphones_Reverberation_Cancellation_for_Distant_Speech_Recognition. Xuewei Zhang.]]<br />
<br />
[[文件:Clipping-speaker.png|200px]]<br />
*[[媒体文件:Clip.pdf|TRP-20150004: Detection and Reconstruction of Clipped Speech in Speaker Recognition. Fanhu Bie et al.]]<br />
<br />
[[文件:Semi-dynamic-embedding.png|200px]]<br />
*[[媒体文件:Taglm.pdf|TRP-20150003: Semi-Dynamic Graph Embedding for Large Scale Language Model Adaptation. Bin Yuan et al.]]<br />
<br />
[[文件:Speaker-discriminative-score.png|200px]]<br />
*[[媒体文件:DNN-based Discriminative Scoring for Speaker.pdf|TRP-20150002: DNN-based Discriminative Scoring for Speaker Recognition Based on i-vector. Jun Wang et al. ]]<br />
<br />
[[文件:Noisy-traiing.png|200px]]<br />
*[[媒体文件:Noisy Training for Deep Neural Networks in.pdf|TRP-20150001: Noisy Training for Deep Neural Networks in Speech Recognition. Shi Yin et al.]]<br />
<br />
[[文件:English-scroing.png|200px]]<br />
*[[媒体文件:AutomaticScoringforEnglishUtterances.pdf|TRP-20140001: Automatic Scoring for English Utterances. Bo Hu.]]<br />
<br />
<br />
[[媒体文件:Template.rar|CSLT-TRP latex template]]</div>Ltyhttp://index.cslt.org/mediawiki/index.php/Publication-trpPublication-trp2016-06-20T19:31:49Z<p>Lty：</p>
<hr />
<div>[[文件:Aikefu.bmp|200px]]<br />
*[[媒体文件:Cslt-trp-template.pdf|TRP-20160004: A Review of Neural QA, Tianyi Luo and Dong Wang]]<br />
<br />
[[文件:simpair.png|200px]]<br />
*[[媒体文件:Trp20160003.pdf|TRP-20160003: A study of Similar Word Model for Unfrequent Word Enhancement in Speech Recognition, Xi Ma, Dong Wang and Javier Tejedor]]<br />
<br />
[[文件:low-freq.png|200px]]<br />
*[[媒体文件:How to deal with low frequency words.pdf|TRP-20160002: Low-Frequency Words Embedding, Chao Xing, Yiqiao Pan, Dong Wang]]<br />
[[文件:maxmargin.png|200px]]<br />
*[[媒体文件:Max-margin.pdf|TRP-20160001: Max-margin metric learning for speaker recognition, Lantian Li, Chao Xing, Dong Wang]]<br />
<br />
<br />
[[文件:lowv.png|200px]]<br />
*[[媒体文件:Lowv.pdf|TRP-20150033: Learning Ordered Word Representations, Xiaoxi Wang, Chao Xing, Dong Wang, Rong Liu and Yiqiao Pan]]<br />
<br />
[[文件:Adamax.png|200px]]<br />
*[[媒体文件:Adamax Online Training for Speech Recognition.pdf|TRP-20150032: Adamax Online Training for Speech Recognition, Xiangyu Zeng, Zhiyong Zhang, Dong Wang]]<br />
<br />
[[文件:Ptrnets.png|200px]]<br />
*[[媒体文件: Ptrnets.pdf|TRP-20150031: An implementation of Pointer-Networks with Extensions, Xiaoxi Wang, Dong Wang]]<br />
<br />
[[文件:dvad.png|200px]]<br />
*[[媒体文件:dvad.pdf|TRP-20150030: DNN-based Voice Activity Detection for Speaker Recognition, Fanhu Bie, Zhiyong Zhang, Dong Wang, Thomas Fang Zheng]]<br />
<br />
[[文件:uyghur.jpg|200px]]<br />
*[[媒体文件:urghur.pdf|TRP-20150029: THUYG-20：A Free Uyghur Speech Database, Askar Rozi, Shi Yin, Zhiyong Zhang, Dong Wang, Askar Hamdulla]]<br />
<br />
[[文件:nnpre.jpg|200px]]<br />
*[[媒体文件:nnpre.pdf|TRP-20150028: Knowledge Transfer Pre-training, Zhiyuang Tang, Dong Wang]]<br />
<br />
[[文件:mmargin.png|200px]]<br />
*[[媒体文件:mmargin.pdf|TRP-20150027: Max-Margin Metric Learning for Speaker Recognition, Lantian Li, Chao Xing, Dong Wang, Thomas Fang Zheng]]<br />
<br />
[[文件:binary.jpg|200px]]<br />
*[[媒体文件:binary.pdf|TRP-20150026: Binary Speaker Embedding, Lantian Li, Chao Xing, Dong Wang, Kaimin Yu, Thomas Fang Zheng]]<br />
<br />
[[文件:rnnrl.png|200px]]<br />
*[[媒体文件:rnnrl.pdf|TRP-20150025: Relation Classification via Recurrent Neural Network, Dong Xu Zhang, Dong Wang, Rong Liu]]<br />
<br />
<br />
[[文件:dplda.png|200px]]<br />
*[[媒体文件:dplda.pdf|TRP-20150024: Learning from LDA using Deep Neural Networks, Dongxu Zhang, Tianyi Luo and Dong Wang]]<br />
<br />
<br />
[[文件:jsrl.png|200px]]<br />
*[[媒体文件:jsrl.pdf|TRP-20150023: Joint Semantic Relevance Learning with Text Data and Graph Knowledge, Dongxu Zhang, Bin Yuan, Dong Wang, Rong Liu]]<br />
<br />
<br />
[[文件:listnet.png|200px]]<br />
*[[媒体文件:listnet.pdf|TRP-20150022: Stochastic Top-k ListNet, Tianyi Luo, Dong Wang, Rong Liu, Yiqiao Pan]]<br />
<br />
<br />
[[文件:segvector.png|200px]]<br />
*[[媒体文件:segvector.pdf|TRP-20150021: Improved Deep Speaker Feature Learning for Text-Dependent Speaker Recognition, Lantian Li, Yiye Lin, Zhiyong Zhang, Dong Wang]]<br />
<br />
<br />
[[文件:Vmclass.png|200px]]<br />
*[[媒体文件:Vmclass.pdf|TRP-20150020: Document Classification with Spherical Word Vectors, Yiqiao Pan, Chao Xing, Dong Wang]]<br />
<br />
<br />
[[文件:Tlearn.png|200px]]<br />
*[[媒体文件:Tlearn.pdf|TRP-20150019: Transfer Learning for Speech and Language Processing, Dong Wang and Thomas Fang Zheng]]<br />
<br />
[[文件:Songcisample.png|200px]]<br />
*[[http://cslt.riit.tsinghua.edu.cn/mediawiki/images/7/7a/Cslt20150018_revisedversion.pdf|TRP-20150018: Chinese Song Iambics Generation with Neural Attention-based Model, Qixin Wang, Tianyi Luo, Dong Wang, Chao Xing]]<br />
<br />
[[文件:database.jpg|200px]]<br />
*[[媒体文件:Thuyg20-sre.pdf|TRP-20150017: AN OPEN/FREE DATABASE AND BENCHMARK FOR UYGHUR SPEAKER RECOGNITION, Askar Rozi, Dong Wang, Zhiyong Zhang, Thomas Fang Zheng]]<br />
<br />
[[文件:Thchs.png|200px]]<br />
*[[媒体文件:Thchs30.pdf|TRP-20150016: THCHS-30 : A Free Chinese Speech Corpus, Dong Wang and Xuewei Zhang]]<br />
<br />
<br />
[[文件:Su.jpg|200px]]<br />
*[[媒体文件:SUSR.pdf|TRP-20150015: Improving Short Utterance Speaker Recognition by Modeling Speech Unit Classes, Chenhao Zhang Dong Wang, Lantian Li and Thomas Fang Zheng]]<br />
<br />
[[文件:dv.png|200px]]<br />
*[[媒体文件:Dvector.pdf|TRP-20150014: Deep Speaker Vectors for Semi Text-independent Speaker Verification, Lantian Li, Dong Wang, Zhiyong Zhang and Thomas Fang Zheng]]<br />
<br />
[[文件:dark.png|200px]]<br />
*[[媒体文件:Dark.pdf|TRP-20150013: Recurrent Neural Network Training with Dark Knowledge Transfer, Dong Wang, Chao Liu, Zhiyuan Tang, Zhiyong Zhang, Mengyuan Zhao]]<br />
<br />
[[文件:PBE.png|200px]]<br />
*[[媒体文件:Probabilistic_Belief_Embedding_for_Knowledge_Population_(TRP).pdf|TRP-20150012: Probabilistic Belief Embedding for Large-scale Knowledge Population. Miao Fan, Qiang Zhou, Andrew Abel, Thomas Fang Zheng and Ralph Grishman]]<br />
<br />
[[文件:fst-fw.png|200px]]<br />
*[[媒体文件:wpair.pdf|TRP-20150011: Recognize Foreign Low-Frequency Words with Similar Pairs, Xi Ma1, Xiaoxi Wang1 and Dong Wang]]<br />
<br />
[[文件:Cdae.png|200px]]<br />
*[[媒体文件:Music.pdf|TRP-20150010: Music Removal by Denoising Autoencoder in Speech Recognition. Mengyuan Zhao, Dong Wang, Zhiyong Zhang and Xuewei Zhang]]<br />
<br />
[[文件:vmfsne.png|200px]]<br />
*[[媒体文件:Cslt-trp-template-vmfsne.pdf|TRP-20150009: VMF-SNE: Embedding for Spherical Data. Mian Wang, Dong Wang]]<br />
<br />
[[文件:ros.png|200px]]<br />
*[[媒体文件:Ros.pdf|TRP-20150008: Learning Speech Rate in Speech Recognition. Xiangyu Zeng, Shi Yin.]]<br />
<br />
[[文件:Dnnvadstru.png|200px]]<br />
*[[媒体文件:DNNVADTRP.pdf|TRP-20150007: Voice Activity Detection Based on Deep Neural Networks. Shi Yin.]] ([[媒体文件:Vad.pdf|Paper submiited to Tsinghua Xuebao]])<br />
<br />
[[文件:Uyghur-training.png|200px]]<br />
*[[媒体文件:UyghurTRP.pdf|TRP-20150006: Low-resource Uyghur Acoustic Model Training based on Cross-lingual Features. Shi Yin.]]<br />
<br />
[[文件:Beam-forming.png|200px]]<br />
*[[媒体文件:Multi-Microphones_Reverberation_Cancellation_for_Distant_Speech_Recognition.pdf|TRP-20150005: Multi-Microphones_Reverberation_Cancellation_for_Distant_Speech_Recognition. Xuewei Zhang.]]<br />
<br />
[[文件:Clipping-speaker.png|200px]]<br />
*[[媒体文件:Clip.pdf|TRP-20150004: Detection and Reconstruction of Clipped Speech in Speaker Recognition. Fanhu Bie et al.]]<br />
<br />
[[文件:Semi-dynamic-embedding.png|200px]]<br />
*[[媒体文件:Taglm.pdf|TRP-20150003: Semi-Dynamic Graph Embedding for Large Scale Language Model Adaptation. Bin Yuan et al.]]<br />
<br />
[[文件:Speaker-discriminative-score.png|200px]]<br />
*[[媒体文件:DNN-based Discriminative Scoring for Speaker.pdf|TRP-20150002: DNN-based Discriminative Scoring for Speaker Recognition Based on i-vector. Jun Wang et al. ]]<br />
<br />
[[文件:Noisy-traiing.png|200px]]<br />
*[[媒体文件:Noisy Training for Deep Neural Networks in.pdf|TRP-20150001: Noisy Training for Deep Neural Networks in Speech Recognition. Shi Yin et al.]]<br />
<br />
[[文件:English-scroing.png|200px]]<br />
*[[媒体文件:AutomaticScoringforEnglishUtterances.pdf|TRP-20140001: Automatic Scoring for English Utterances. Bo Hu.]]<br />
<br />
<br />
[[媒体文件:Template.rar|CSLT-TRP latex template]]</div>Ltyhttp://index.cslt.org/mediawiki/index.php/Publication-trpPublication-trp2016-06-20T19:30:35Z<p>Lty：</p>
<hr />
<div>[[文件:Aikefu.bmp|200px]]<br />
*[[媒体文件:Cslt-trp-template.pdf|TRP-20160004: A Review of Neural QA, Tianyi Luo and Dong Wang]]<br />
<br />
[[文件:simpair.png|200px]]<br />
*[[媒体文件:Trp20160003.pdf|TRP-20160003: A study of Similar Word Model for Unfrequent Word Enhancement in Speech Recognition, Xi Ma, Dong Wang and Javier Tejedor]]<br />
<br />
[[文件:low-freq.png|200px]]<br />
*[[媒体文件:How to deal with low frequency words.pdf|TRP-20160002: Low-Frequency Words Embedding, Chao Xing, Yiqiao Pan, Dong Wang]]<br />
[[文件:maxmargin.png|200px]]<br />
*[[媒体文件:Max-margin.pdf|TRP-20160001: Max-margin metric learning for speaker recognition, Lantian Li, Chao Xing, Dong Wang]]<br />
<br />
<br />
[[文件:lowv.png|200px]]<br />
*[[媒体文件:Lowv.pdf|TRP-20150033: Learning Ordered Word Representations, Xiaoxi Wang, Chao Xing, Dong Wang, Rong Liu and Yiqiao Pan]]<br />
<br />
[[文件:Adamax.png|200px]]<br />
*[[媒体文件:Adamax Online Training for Speech Recognition.pdf|TRP-20150032: Adamax Online Training for Speech Recognition, Xiangyu Zeng, Zhiyong Zhang, Dong Wang]]<br />
<br />
[[文件:Ptrnets.png|200px]]<br />
*[[媒体文件: Ptrnets.pdf|TRP-20150031: An implementation of Pointer-Networks with Extensions, Xiaoxi Wang, Dong Wang]]<br />
<br />
[[文件:dvad.png|200px]]<br />
*[[媒体文件:dvad.pdf|TRP-20150030: DNN-based Voice Activity Detection for Speaker Recognition, Fanhu Bie, Zhiyong Zhang, Dong Wang, Thomas Fang Zheng]]<br />
<br />
[[文件:uyghur.jpg|200px]]<br />
*[[媒体文件:urghur.pdf|TRP-20150029: THUYG-20：A Free Uyghur Speech Database, Askar Rozi, Shi Yin, Zhiyong Zhang, Dong Wang, Askar Hamdulla]]<br />
<br />
[[文件:nnpre.jpg|200px]]<br />
*[[媒体文件:nnpre.pdf|TRP-20150028: Knowledge Transfer Pre-training, Zhiyuang Tang, Dong Wang]]<br />
<br />
[[文件:mmargin.png|200px]]<br />
*[[媒体文件:mmargin.pdf|TRP-20150027: Max-Margin Metric Learning for Speaker Recognition, Lantian Li, Chao Xing, Dong Wang, Thomas Fang Zheng]]<br />
<br />
[[文件:binary.jpg|200px]]<br />
*[[媒体文件:binary.pdf|TRP-20150026: Binary Speaker Embedding, Lantian Li, Chao Xing, Dong Wang, Kaimin Yu, Thomas Fang Zheng]]<br />
<br />
[[文件:rnnrl.png|200px]]<br />
*[[媒体文件:rnnrl.pdf|TRP-20150025: Relation Classification via Recurrent Neural Network, Dong Xu Zhang, Dong Wang, Rong Liu]]<br />
<br />
<br />
[[文件:dplda.png|200px]]<br />
*[[媒体文件:dplda.pdf|TRP-20150024: Learning from LDA using Deep Neural Networks, Dongxu Zhang, Tianyi Luo and Dong Wang]]<br />
<br />
<br />
[[文件:jsrl.png|200px]]<br />
*[[媒体文件:jsrl.pdf|TRP-20150023: Joint Semantic Relevance Learning with Text Data and Graph Knowledge, Dongxu Zhang, Bin Yuan, Dong Wang, Rong Liu]]<br />
<br />
<br />
[[文件:listnet.png|200px]]<br />
*[[媒体文件:listnet.pdf|TRP-20150022: Stochastic Top-k ListNet, Tianyi Luo, Dong Wang, Rong Liu, Yiqiao Pan]]<br />
<br />
<br />
[[文件:segvector.png|200px]]<br />
*[[媒体文件:segvector.pdf|TRP-20150021: Improved Deep Speaker Feature Learning for Text-Dependent Speaker Recognition, Lantian Li, Yiye Lin, Zhiyong Zhang, Dong Wang]]<br />
<br />
<br />
[[文件:Vmclass.png|200px]]<br />
*[[媒体文件:Vmclass.pdf|TRP-20150020: Document Classification with Spherical Word Vectors, Yiqiao Pan, Chao Xing, Dong Wang]]<br />
<br />
<br />
[[文件:Tlearn.png|200px]]<br />
*[[媒体文件:Tlearn.pdf|TRP-20150019: Transfer Learning for Speech and Language Processing, Dong Wang and Thomas Fang Zheng]]<br />
<br />
[[文件:Songcisample.png|200px]]<br />
*[[媒体文件:http://cslt.riit.tsinghua.edu.cn/mediawiki/images/7/7a/Cslt20150018_revisedversion.pdf|TRP-20150018: Chinese Song Iambics Generation with Neural Attention-based Model, Qixin Wang, Tianyi Luo, Dong Wang, Chao Xing]]<br />
<br />
[[文件:database.jpg|200px]]<br />
*[[媒体文件:Thuyg20-sre.pdf|TRP-20150017: AN OPEN/FREE DATABASE AND BENCHMARK FOR UYGHUR SPEAKER RECOGNITION, Askar Rozi, Dong Wang, Zhiyong Zhang, Thomas Fang Zheng]]<br />
<br />
[[文件:Thchs.png|200px]]<br />
*[[媒体文件:Thchs30.pdf|TRP-20150016: THCHS-30 : A Free Chinese Speech Corpus, Dong Wang and Xuewei Zhang]]<br />
<br />
<br />
[[文件:Su.jpg|200px]]<br />
*[[媒体文件:SUSR.pdf|TRP-20150015: Improving Short Utterance Speaker Recognition by Modeling Speech Unit Classes, Chenhao Zhang Dong Wang, Lantian Li and Thomas Fang Zheng]]<br />
<br />
[[文件:dv.png|200px]]<br />
*[[媒体文件:Dvector.pdf|TRP-20150014: Deep Speaker Vectors for Semi Text-independent Speaker Verification, Lantian Li, Dong Wang, Zhiyong Zhang and Thomas Fang Zheng]]<br />
<br />
[[文件:dark.png|200px]]<br />
*[[媒体文件:Dark.pdf|TRP-20150013: Recurrent Neural Network Training with Dark Knowledge Transfer, Dong Wang, Chao Liu, Zhiyuan Tang, Zhiyong Zhang, Mengyuan Zhao]]<br />
<br />
[[文件:PBE.png|200px]]<br />
*[[媒体文件:Probabilistic_Belief_Embedding_for_Knowledge_Population_(TRP).pdf|TRP-20150012: Probabilistic Belief Embedding for Large-scale Knowledge Population. Miao Fan, Qiang Zhou, Andrew Abel, Thomas Fang Zheng and Ralph Grishman]]<br />
<br />
[[文件:fst-fw.png|200px]]<br />
*[[媒体文件:wpair.pdf|TRP-20150011: Recognize Foreign Low-Frequency Words with Similar Pairs, Xi Ma1, Xiaoxi Wang1 and Dong Wang]]<br />
<br />
[[文件:Cdae.png|200px]]<br />
*[[媒体文件:Music.pdf|TRP-20150010: Music Removal by Denoising Autoencoder in Speech Recognition. Mengyuan Zhao, Dong Wang, Zhiyong Zhang and Xuewei Zhang]]<br />
<br />
[[文件:vmfsne.png|200px]]<br />
*[[媒体文件:Cslt-trp-template-vmfsne.pdf|TRP-20150009: VMF-SNE: Embedding for Spherical Data. Mian Wang, Dong Wang]]<br />
<br />
[[文件:ros.png|200px]]<br />
*[[媒体文件:Ros.pdf|TRP-20150008: Learning Speech Rate in Speech Recognition. Xiangyu Zeng, Shi Yin.]]<br />
<br />
[[文件:Dnnvadstru.png|200px]]<br />
*[[媒体文件:DNNVADTRP.pdf|TRP-20150007: Voice Activity Detection Based on Deep Neural Networks. Shi Yin.]] ([[媒体文件:Vad.pdf|Paper submiited to Tsinghua Xuebao]])<br />
<br />
[[文件:Uyghur-training.png|200px]]<br />
*[[媒体文件:UyghurTRP.pdf|TRP-20150006: Low-resource Uyghur Acoustic Model Training based on Cross-lingual Features. Shi Yin.]]<br />
<br />
[[文件:Beam-forming.png|200px]]<br />
*[[媒体文件:Multi-Microphones_Reverberation_Cancellation_for_Distant_Speech_Recognition.pdf|TRP-20150005: Multi-Microphones_Reverberation_Cancellation_for_Distant_Speech_Recognition. Xuewei Zhang.]]<br />
<br />
[[文件:Clipping-speaker.png|200px]]<br />
*[[媒体文件:Clip.pdf|TRP-20150004: Detection and Reconstruction of Clipped Speech in Speaker Recognition. Fanhu Bie et al.]]<br />
<br />
[[文件:Semi-dynamic-embedding.png|200px]]<br />
*[[媒体文件:Taglm.pdf|TRP-20150003: Semi-Dynamic Graph Embedding for Large Scale Language Model Adaptation. Bin Yuan et al.]]<br />
<br />
[[文件:Speaker-discriminative-score.png|200px]]<br />
*[[媒体文件:DNN-based Discriminative Scoring for Speaker.pdf|TRP-20150002: DNN-based Discriminative Scoring for Speaker Recognition Based on i-vector. Jun Wang et al. ]]<br />
<br />
[[文件:Noisy-traiing.png|200px]]<br />
*[[媒体文件:Noisy Training for Deep Neural Networks in.pdf|TRP-20150001: Noisy Training for Deep Neural Networks in Speech Recognition. Shi Yin et al.]]<br />
<br />
[[文件:English-scroing.png|200px]]<br />
*[[媒体文件:AutomaticScoringforEnglishUtterances.pdf|TRP-20140001: Automatic Scoring for English Utterances. Bo Hu.]]<br />
<br />
<br />
[[媒体文件:Template.rar|CSLT-TRP latex template]]</div>Ltyhttp://index.cslt.org/mediawiki/index.php/Conference_AgendaConference Agenda2016-06-19T00:36:41Z<p>Lty：</p>
<hr />
<div>{| class="wikitable"<br />
!Conference !! Remaining days !! Venue !! submission deadline !! conference date !! target people <br />
|-<br />
|ICASSP 2016 || Pass ||Shanghai, China || 9/25/2015 || 3/20/2016-3/25/2016 || WD<br />
|-<br />
|NAACL 2016 || Pass ||San Diego, CA || 1/6/2015 || 6/13/2016-6/15/2016 || ZDX<br />
|-<br />
|IJCAI 2016 || Pass ||New York, NY || 1/27/2016(Abstract);2/2/2016(papers) || 7/9/2016-7/13/2016 || LTY, WQX<br />
|-<br />
|ACL 2016 || Pass || Berlin, Germany || 2/29/2016(short) 3/18/2016(long) || 8/7/2016–8/12/2016 || LTY, WQX<br />
|-<br />
|Interspeech 2016 || Pass ||San Francisco, CA|| 3/30/2016 || 9/8/2016-9/12/2016 ||<br />
|-<br />
|ICDM 2016 || Pass ||Venice, Italy || 5/7/2016 || 11/7/2016-11/8/2016 || <br />
|-<br />
|NIPS 2016 || Pass ||Barcelona Spain || 5/20/2016 || 11/7/2016-11/8/2016 || <br />
|-<br />
|APSIPA 2016 || Pass ||Jeju, Korea || 6/15/2016 || 12/13/2016-12/16/2016 || <br />
|-<br />
|CCL 2016 || Pass ||Qingdao, China|| 6/1/2016 || 10/15/2016-10/16/2016 || LTY<br />
|-<br />
|EMNLP 2016 || Pass ||Austin, TX || 6/3/2016 || 12/5/2016-12/10/2016 || <br />
|-<br />
|ISCSLP 2016 || Pass ||Tianjin, CN || 6/17/2016 || 10/17/2016-10/20/2016 || <br />
|-<br />
|BIC 2016 || Pass ||Beijing, CN || 6/10/2016 || 11/28/2016-11/30/2016 || <br />
|-<br />
|OCOCOSDA 2016 || 11 days ||Baili || 6/29/2016 || 10/26/2016-10/28/2016 || <br />
|-<br />
|COLING 2016 || 27 days ||Osaka, Japan || 7/15/2016 || 12/11/2016-12/16/2016 || <br />
|-<br />
|}<br />
<br />
<br />
[http://www.cs.rochester.edu/~tetreaul/conferences.html NLP conference list from Joel Tetreault]<br />
<br />
[http://www.aclclp.org.tw/confer_c.php CFP list from ACLCLP]<br />
<br />
<br />
[[past-conf-2014|2014]]<br />
<br />
[[past-conf-2015|2015]]</div>Ltyhttp://index.cslt.org/mediawiki/index.php/Tianyi_Luo_2016-05-09Tianyi Luo 2016-05-092016-05-09T10:46:34Z<p>Lty：</p>
<hr />
<div>=== Plan to do this week ===<br />
* To implement tensorflow version of RNN/LSTM Max margin vector training.<br />
* To implement attention chatting model with xiaobing corpus.<br />
=== Work done in this week ===<br />
--------------------2016-05-02~05-04<br />
* The Holiday of Labors Day.<br />
--------------------2016-05-05<br />
* Finish the preprocessing of Xiaobing corpus.<br />
* Implement part of code about the qqa (Current sample is q1; Positive sample is a1; Negtive sample is q2.) max-margin theano version.<br />
--------------------2016-05-06<br />
* Finish implementing code about the qqa (Current sample is q1; Positive sample is a1; Negtive sample is q2.) max-margin theano version.<br />
--------------------2016-05-07<br />
* Wait for the results of experiments.<br />
* Ready for going to Silicon Valley.<br />
--------------------2016-05-08<br />
* Arrived in Silicon Valley.<br />
<br />
=== Plan to do next week ===<br />
* To implement tensorflow version of RNN/LSTM Max margin vector training.<br />
* To implement attention chatting model with xiaobing corpus.<br />
===Interested papers ===<br />
*Cascading Bandits: Learning to Rank in the Cascade Model(ICML 2015) [[http://zheng-wen.com/Cascading_Bandit_Paper.pdf pdf]]</div>Ltyhttp://index.cslt.org/mediawiki/index.php/Tianyi_Luo_2016-05-02Tianyi Luo 2016-05-022016-05-09T10:44:08Z<p>Lty：</p>
<hr />
<div>=== Plan to do this week ===<br />
* To implement tensorflow version of RNN/LSTM Max margin vector training.<br />
* To implement attention chatting model with xiaobing corpus.<br />
=== Work done in this week ===<br />
--------------------2016-04-25<br />
* Conduct some preprocessing of xiaobing corpus.<br />
--------------------2016-04-26<br />
* Help jiyuan to understand the music generation paper.<br />
* Check jiyuan's code(from RNN-RBM to LSTM-RBM).<br />
--------------------2016-04-27<br />
* Modify the LSTM Max margin vector training code. Its cost is lower than RNN and the training speed is slower than RNN. The performance improves from 82.39% to 84.83%.<br />
--------------------2016-04-28<br />
* Finish part of code about qaa(Current sample is q1; Positive sample is a1; Negtive sampling is a2.) max margin theano version.<br />
--------------------2016-04-29<br />
* Finish the code about qaa(Current sample is q1; Positive sample is a1; Negtive sampling is a2.) max margin theano version.<br />
--------------------2016-04-30~2016-05-01<br />
The Holiday of Labors Day<br />
<br />
<br />
=== Plan to do next week ===<br />
* To implement tensorflow version of RNN/LSTM Max margin vector training.<br />
* To implement attention chatting model with xiaobing corpus.<br />
===Interested papers ===<br />
*Cascading Bandits: Learning to Rank in the Cascade Model(ICML 2015) [[http://zheng-wen.com/Cascading_Bandit_Paper.pdf pdf]]</div>Ltyhttp://index.cslt.org/mediawiki/index.php/Tianyi_Luo_2016-05-09Tianyi Luo 2016-05-092016-05-09T10:39:49Z<p>Lty：以“=== Plan to do this week === * To implement tensorflow version of RNN/LSTM Max margin vector training. * To implement attention chatting model with xiaobing corpus....”为内容创建页面</p>
<hr />
<div>=== Plan to do this week ===<br />
* To implement tensorflow version of RNN/LSTM Max margin vector training.<br />
* To implement attention chatting model with xiaobing corpus.<br />
=== Work done in this week ===<br />
--------------------2016-05-02~05-04<br />
* The Holiday of Labors Day.<br />
--------------------2016-05-05<br />
* Finish the preprocessing of Xiaobing corpus.<br />
* Implement the max-margin theano version(Sample training pair is: q1,a1,q2. It means the negative sample is q.).<br />
--------------------2016-05-06<br />
* Implement the max-margin theano version(Sample training pair is: q1,a1,a2. It means the negative sample is a.).<br />
--------------------2016-05-07<br />
* Wait for the results of experiments.<br />
* Ready for going to Silicon Valley.<br />
--------------------2016-05-08<br />
* Arrived in Silicon Valley.<br />
<br />
=== Plan to do next week ===<br />
* To implement tensorflow version of RNN/LSTM Max margin vector training.<br />
* To implement attention chatting model with xiaobing corpus.<br />
===Interested papers ===<br />
*Cascading Bandits: Learning to Rank in the Cascade Model(ICML 2015) [[http://zheng-wen.com/Cascading_Bandit_Paper.pdf pdf]]</div>Ltyhttp://index.cslt.org/mediawiki/index.php/2016-05-092016-05-092016-05-09T10:23:34Z<p>Lty：</p>
<hr />
<div>[[Mengyuan Zhao 2016-05-09]]<br />
<br />
[[Lantian Li 2016-05-09]]<br />
<br />
[[Yang Wang 2016-05-09]]<br />
<br />
[[Xuewei Zhang 2016-05-09]]<br />
<br />
[[Zhiyong Zhang 2016-05-09]]<br />
<br />
[[Chao Xing 2016-05-09]]<br />
<br />
[[Zhiyuan Tang 2016-05-09]]<br />
<br />
[[Aodong Li 2016-05-09]]<br />
<br />
[[jiyuan Zhang 2016-05-09]]<br />
<br />
[[Yiqiao Pan 2016-05-09]]<br />
<br />
[[Xiangyu Zeng 2016-05-09]]<br />
<br />
[[Tianyi Luo 2016-05-09]]</div>Ltyhttp://index.cslt.org/mediawiki/index.php/Tianyi_Luo_2016-05-02Tianyi Luo 2016-05-022016-04-28T09:13:25Z<p>Lty：</p>
<hr />
<div>=== Plan to do this week ===<br />
* To implement tensorflow version of RNN/LSTM Max margin vector training.<br />
* To implement attention chatting model with xiaobing corpus.<br />
=== Work done in this week ===<br />
--------------------2016-04-25<br />
* Conduct some preprocess of xiaobing corpus.<br />
--------------------2016-04-26<br />
* Help jiyuan to understand the music generation paper.<br />
* Check jiyuan's code(from RNN-RBM to LSTM-RBM).<br />
--------------------2016-04-27<br />
* Modify the LSTM Max margin vector training code. Its cost is lower than RNN and the training speed is slower than RNN. The performance improves from 82.39% to 84.83%.<br />
--------------------2016-04-28<br />
<br />
--------------------2016-04-29<br />
<br />
--------------------2016-04-30<br />
<br />
--------------------2016-05-01<br />
<br />
<br />
=== Plan to do next week ===<br />
* To implement tensorflow version of RNN/LSTM Max margin vector training.<br />
* To implement attention chatting model with xiaobing corpus.<br />
===Interested papers ===<br />
*Cascading Bandits: Learning to Rank in the Cascade Model(ICML 2015) [[http://zheng-wen.com/Cascading_Bandit_Paper.pdf pdf]]</div>Ltyhttp://index.cslt.org/mediawiki/index.php/Tianyi_Luo_2016-05-02Tianyi Luo 2016-05-022016-04-28T06:26:44Z<p>Lty：</p>
<hr />
<div>=== Plan to do this week ===<br />
* To implement tensorflow version of RNN/LSTM Max margin vector training.<br />
* To implement attention chatting model with xiaobing corpus.<br />
=== Work done in this week ===<br />
--------------------2016-04-25<br />
* Conduct some preprocess of xiaobing corpus.<br />
--------------------2016-04-26<br />
* Help jiyuan to understand the music generation paper.<br />
* Check jiyuan's code(from RNN-RBM to LSTM-RBM).<br />
--------------------2016-04-27<br />
* Modify the LSTM Max margin vector training. Its cost is lower than RNN and the training speed is slower than RNN. The performance improve from 82.39% to 84.83%.<br />
--------------------2016-04-28<br />
<br />
--------------------2016-04-29<br />
<br />
--------------------2016-04-30<br />
<br />
--------------------2016-05-01<br />
<br />
<br />
=== Plan to do next week ===<br />
* To implement tensorflow version of RNN/LSTM Max margin vector training.<br />
* To implement attention chatting model with xiaobing corpus.<br />
===Interested papers ===<br />
*Cascading Bandits: Learning to Rank in the Cascade Model(ICML 2015) [[http://zheng-wen.com/Cascading_Bandit_Paper.pdf pdf]]</div>Ltyhttp://index.cslt.org/mediawiki/index.php/Tianyi_Luo_2016-05-02Tianyi Luo 2016-05-022016-04-27T18:52:03Z<p>Lty：</p>
<hr />
<div>=== Plan to do this week ===<br />
* To implement tensorflow version of RNN/LSTM Max margin vector training.<br />
* To implement attention chatting model with xiaobing corpus.<br />
=== Work done in this week ===<br />
--------------------2016-04-25<br />
* Conduct some preprocess of xiaobing corpus.<br />
--------------------2016-04-26<br />
* Help jiyuan to understand the music generation paper.<br />
* Check jiyuan's code(from RNN-RBM to LSTM-RBM).<br />
--------------------2016-04-27<br />
* Modify the LSTM Max margin vector training. Its cost is lower than RNN and the training speed is slower than RNN.<br />
--------------------2016-04-28<br />
<br />
--------------------2016-04-29<br />
<br />
--------------------2016-04-30<br />
<br />
--------------------2016-05-01<br />
<br />
<br />
=== Plan to do next week ===<br />
* To implement tensorflow version of RNN/LSTM Max margin vector training.<br />
* To implement attention chatting model with xiaobing corpus.<br />
===Interested papers ===<br />
*Cascading Bandits: Learning to Rank in the Cascade Model(ICML 2015) [[http://zheng-wen.com/Cascading_Bandit_Paper.pdf pdf]]</div>Ltyhttp://index.cslt.org/mediawiki/index.php/Tianyi_Luo_2016-05-02Tianyi Luo 2016-05-022016-04-27T18:50:58Z<p>Lty：</p>
<hr />
<div>=== Plan to do this week ===<br />
* To implement tensorflow version of RNN/LSTM Max margin vector training.<br />
* To implement attention chatting model with xiaobing corpus.<br />
=== Work done in this week ===<br />
--------------------2016-04-25<br />
--------------------2016-04-26<br />
* Help jiyuan to understand the music generation paper.<br />
* Check jiyuan's code(from RNN-RBM to LSTM-RBM).<br />
--------------------2016-04-27<br />
* Modify the LSTM Max margin vector training. Its cost is lower than RNN and the training speed is slower than RNN.<br />
--------------------2016-04-28<br />
<br />
--------------------2016-04-29<br />
<br />
--------------------2016-04-30<br />
<br />
--------------------2016-05-01<br />
<br />
<br />
=== Plan to do next week ===<br />
* To implement tensorflow version of RNN/LSTM Max margin vector training.<br />
* To implement attention chatting model with xiaobing corpus.<br />
===Interested papers ===<br />
*Cascading Bandits: Learning to Rank in the Cascade Model(ICML 2015) [[http://zheng-wen.com/Cascading_Bandit_Paper.pdf pdf]]</div>Ltyhttp://index.cslt.org/mediawiki/index.php/Tianyi_Luo_2016-05-02Tianyi Luo 2016-05-022016-04-27T18:49:10Z<p>Lty：</p>
<hr />
<div>=== Plan to do this week ===<br />
* To implement tensorflow version of RNN/LSTM Max margin vector training.<br />
* To implement attention chatting model with xiaobing corpus.<br />
=== Work done in this week ===<br />
--------------------2016-04-25<br />
<br />
--------------------2016-04-26<br />
<br />
--------------------2016-04-27<br />
* Implement LSTM Max margin vector training. Its cost is lower than RNN and the training speed is slower than RNN.<br />
--------------------2016-04-28<br />
<br />
--------------------2016-04-29<br />
<br />
--------------------2016-04-30<br />
<br />
--------------------2016-05-01<br />
<br />
<br />
=== Plan to do next week ===<br />
* To implement tensorflow version of RNN/LSTM Max margin vector training.<br />
* To implement attention chatting model with xiaobing corpus.<br />
===Interested papers ===<br />
*Cascading Bandits: Learning to Rank in the Cascade Model(ICML 2015) [[http://zheng-wen.com/Cascading_Bandit_Paper.pdf pdf]]</div>Ltyhttp://index.cslt.org/mediawiki/index.php/Tianyi_Luo_2016-05-02Tianyi Luo 2016-05-022016-04-27T18:47:33Z<p>Lty：</p>
<hr />
<div>=== Plan to do this week ===<br />
* To implement tensorflow version of RNN/LSTM Max margin vector training.<br />
* To implement attention chatting model with xiaobing corpus.<br />
=== Work done in this week ===<br />
--------------------2016-04-25<br />
<br />
--------------------2016-04-26<br />
<br />
--------------------2016-04-27<br />
<br />
--------------------2016-04-28<br />
<br />
--------------------2016-04-29<br />
<br />
--------------------2016-04-30<br />
<br />
--------------------2016-05-01<br />
<br />
<br />
=== Plan to do next week ===<br />
* To implement tensorflow version of RNN/LSTM Max margin vector training.<br />
* To implement attention chatting model with xiaobing corpus.<br />
===Interested papers ===<br />
*Cascading Bandits: Learning to Rank in the Cascade Model(ICML 2015) [[http://zheng-wen.com/Cascading_Bandit_Paper.pdf pdf]]</div>Ltyhttp://index.cslt.org/mediawiki/index.php/Tianyi_Luo_2016-05-02Tianyi Luo 2016-05-022016-04-27T18:46:49Z<p>Lty：以“=== Plan to do this week === * To implement tensorflow version of RNN/LSTM Max margin vector training. === Work done in this week === --------------------2016-04-26...”为内容创建页面</p>
<hr />
<div>=== Plan to do this week ===<br />
* To implement tensorflow version of RNN/LSTM Max margin vector training.<br />
=== Work done in this week ===<br />
--------------------2016-04-26<br />
<br />
--------------------2016-04-27<br />
<br />
--------------------2016-04-28<br />
<br />
--------------------2016-04-29<br />
<br />
--------------------2016-04-30<br />
<br />
--------------------2016-05-01<br />
<br />
--------------------2016-05-02<br />
<br />
<br />
=== Plan to do next week ===<br />
* To implement tensorflow version of RNN/LSTM Max margin vector training.<br />
* To implement attention chatting model with xiaobing corpus.<br />
===Interested papers ===<br />
*Cascading Bandits: Learning to Rank in the Cascade Model(ICML 2015) [[http://zheng-wen.com/Cascading_Bandit_Paper.pdf pdf]]</div>Ltyhttp://index.cslt.org/mediawiki/index.php/2016-05-022016-05-022016-04-27T18:45:12Z<p>Lty：以“Tianyi Luo 2016-05-02”为内容创建页面</p>
<hr />
<div>[[Tianyi Luo 2016-05-02]]</div>Ltyhttp://index.cslt.org/mediawiki/index.php/Status_reportStatus report2016-04-27T18:44:46Z<p>Lty：</p>
<hr />
<div><br />
[[2016-01-04]]<br />
<br />
[[2016-01-11]]<br />
<br />
[[2016-01-18]]<br />
<br />
[[2016-01-25]]<br />
<br />
[[2016-02-01]]<br />
<br />
[[2016-02-22]]<br />
<br />
[[2016-02-29]]<br />
<br />
[[2016-03-07]]<br />
<br />
[[2016-03-14]]<br />
<br />
[[2016-03-21]]<br />
<br />
[[2016-03-28]]<br />
<br />
[[2016-04-04]]<br />
<br />
[[2016-04-11]]<br />
<br />
[[2016-04-18]]<br />
<br />
[[2016-04-25]]<br />
<br />
[[2016-05-02]]</div>Ltyhttp://index.cslt.org/mediawiki/index.php/Tianyi_Luo_2016-04-25Tianyi Luo 2016-04-252016-04-25T00:36:46Z<p>Lty：</p>
<hr />
<div>=== Plan to do this week ===<br />
* To implement tensorflow version of RNN/LSTM Max margin vector training.<br />
=== Work done in this week ===<br />
--------------------2016-04-18<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
* Finish implementing theano version of LSTM Max margin vector training.<br />
--------------------2016-04-19<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
--------------------2016-04-20<br />
* Finish submiting the camera version paper of IJCAI 2016.<br />
* Update the version of Technical Report about Chinese Song Iambics generation.<br />
--------------------2016-04-21<br />
* Finish helping Teacher Wang to prepare for text group's presentation(Tang poetry and Songci generation and Intelligent QA system) for Tsinghua University's 105 anniversary.<br />
* Submit our IJCAI paper to arxiv. (Solve a big problem about submitting the paper including Chinese chacracters. [http://cslt.riit.tsinghua.edu.cn/mediawiki/index.php/How_to_submit_the_latex_files_including_Chinese_characters_to_arxiv Solution])<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
--------------------2016-04-22<br />
* Speed up process of the test performance about theano version of Generationg the similar questions' vectors based on RNN.<br />
* Help conduct the presentation of text group(Tang poetry and Songci generation and Intelligent QA system) for Tsinghua University's 105 anniversary.<br />
--------------------2016-04-23<br />
* Use entity match rule 1(dpk, ix, lpk sv + machine number) to improve the accuracy from 38% to 58%.<br />
--------------------2016-04-24<br />
* Use entity match rules 2+3(dps, s, e, fp, tps, q, 200 + machine number and machine number only) to improve the accuracy from 58% to 72.09%.<br />
* Use entity match rules 4(dpk, ix, lpk sv, dps, s, e, fp, tps, q + machine number + English chacracters) to improve the accuracy from 72.09% to 72.86%.<br />
* Use entity match rules 5(machine number + English chacracters) to improve the accuracy from 72.86% to 76.48%.<br />
* Use simlar pair match rules 1(“快”和“速度”、“复写”和“拷贝”等等) to improve the accuracy from 76.48% to 82.39%.<br />
=== Plan to do next week ===<br />
* To implement tensorflow version of RNN/LSTM Max margin vector training.<br />
* To implement attention chatting model with xiaobing corpus.<br />
===Interested papers ===<br />
*Cascading Bandits: Learning to Rank in the Cascade Model(ICML 2015) [[http://zheng-wen.com/Cascading_Bandit_Paper.pdf pdf]]</div>Ltyhttp://index.cslt.org/mediawiki/index.php/Conference_AgendaConference Agenda2016-04-24T16:34:30Z<p>Lty：</p>
<hr />
<div>{| class="wikitable"<br />
!Conference !! Remaining days !! Venue !! submission deadline !! conference date !! target people <br />
|-<br />
|ICASSP 2016 || Pass ||Shanghai, China || 9/25/2015 || 3/20/2016-3/25/2016 || WD<br />
|-<br />
|NAACL 2016 || Pass ||San Diego, CA || 1/6/2015 || 6/13/2016-6/15/2016 || ZDX<br />
|-<br />
|IJCAI 2016 || Pass ||New York, NY || 1/27/2016(Abstract);2/2/2016(papers) || 7/9/2016-7/13/2016 || LTY, WQX<br />
|-<br />
|ACL 2016 || Pass || Berlin, Germany || 2/29/2016(short) 3/18/2016(long) || 8/7/2016–8/12/2016 || LTY, WQX<br />
|-<br />
|Interspeech 2016 || Pass ||San Francisco, CA|| 3/30/2016 || 9/8/2016-9/12/2016 ||<br />
|-<br />
|ICDM 2016 || 12 days ||Venice, Italy || 5/7/2016 || 11/7/2016-11/8/2016 || LTY<br />
|-<br />
|NIPS 2016 || 22 days ||Barcelona Spain || 5/20/2016 || 11/7/2016-11/8/2016 || LTY<br />
|-<br />
|APSIPA 2016 || ||Jeju, Korea || 5/31/2016 || 12/13/2016-12/16/2016 || <br />
|-<br />
|CCL 2016 || ||Qingdao, China|| 6/1/2016 || 10/15/2016-10/16/2016 || <br />
|-<br />
|EMNLP 2016 || 41 days ||Austin, TX || 6/3/2016 || 12/5/2016-12/10/2016 || LTY<br />
|-<br />
|ISCSLP 2016 || ||Tianjin, CN || 6/3/2016 || 10/17/2016-10/20/2016 || <br />
|-<br />
|OCOCOSDA 2016 || ||Baili || 6/29/2016 || 10/26/2016-10/28/2016 || <br />
|-<br />
|COLING 2016 || 82 days ||Osaka, Japan || 7/15/2016 || 12/11/2016-12/16/2016 || <br />
|-<br />
|}<br />
<br />
<br />
[http://www.cs.rochester.edu/~tetreaul/conferences.html NLP conference list from Joel Tetreault]<br />
<br />
[http://www.aclclp.org.tw/confer_c.php CFP list from ACLCLP]<br />
<br />
<br />
[[past-conf-2014|2014]]<br />
<br />
[[past-conf-2015|2015]]</div>Ltyhttp://index.cslt.org/mediawiki/index.php/Tianyi_Luo_2016-04-25Tianyi Luo 2016-04-252016-04-24T16:20:30Z<p>Lty：</p>
<hr />
<div>=== Plan to do this week ===<br />
* To implement tensorflow version of RNN/LSTM Max margin vector training.<br />
=== Work done in this week ===<br />
--------------------2016-04-18<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
* Finish implementing theano version of LSTM Max margin vector training.<br />
--------------------2016-04-19<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
--------------------2016-04-20<br />
* Finish submiting the camera version paper of IJCAI 2016.<br />
* Update the version of Technical Report about Chinese Song Iambics generation.<br />
--------------------2016-04-21<br />
* Finish helping Teacher Wang to prepare for text group's presentation(Tang poetry and Songci generation and Intelligent QA system) for Tsinghua University's 105 anniversary.<br />
* Submit our IJCAI paper to arxiv. (Solve a big problem about submitting the paper including Chinese chacracters. [http://cslt.riit.tsinghua.edu.cn/mediawiki/index.php/How_to_submit_the_latex_files_including_Chinese_characters_to_arxiv Solution])<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
--------------------2016-04-22<br />
* Speed up process of the test performance about theano version of Generationg the similar questions' vectors based on RNN.<br />
* Help conduct the presentation of text group(Tang poetry and Songci generation and Intelligent QA system) for Tsinghua University's 105 anniversary.<br />
--------------------2016-04-23<br />
* Use entity match rule 1(dpk, ix, lpk sv + machine number) to improve the accuracy from 38% to 58%.<br />
--------------------2016-04-24<br />
* Use entity match rules 2+3(dps, s, e, fp, tps, q, 200 + machine number and machine number only) to improve the accuracy from 58% to 72.09%.<br />
* Use entity match rules 4(dpk, ix, lpk sv, dps, s, e, fp, tps, q + machine number + English chacracters) to improve the accuracy from 72.09% to 72.86%.<br />
* Use entity match rules 5(machine number + English chacracters) to improve the accuracy from 72.86% to 76.48%.<br />
* Use simlar pair match rules 1(“快”和“速度”、“复写”和“拷贝”等等) to improve the accuracy from 76.48% to ？%.<br />
=== Plan to do next week ===<br />
* To implement tensorflow version of RNN/LSTM Max margin vector training.<br />
* To implement attention chatting model with xiaobing corpus.<br />
===Interested papers ===<br />
*Cascading Bandits: Learning to Rank in the Cascade Model(ICML 2015) [[http://zheng-wen.com/Cascading_Bandit_Paper.pdf pdf]]</div>Ltyhttp://index.cslt.org/mediawiki/index.php/Tianyi_Luo_2016-04-25Tianyi Luo 2016-04-252016-04-24T15:56:26Z<p>Lty：</p>
<hr />
<div>=== Plan to do this week ===<br />
* To implement tensorflow version of RNN/LSTM Max margin vector training.<br />
=== Work done in this week ===<br />
--------------------2016-04-18<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
* Finish implementing theano version of LSTM Max margin vector training.<br />
--------------------2016-04-19<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
--------------------2016-04-20<br />
* Finish submiting the camera version paper of IJCAI 2016.<br />
* Update the version of Technical Report about Chinese Song Iambics generation.<br />
--------------------2016-04-21<br />
* Finish helping Teacher Wang to prepare for text group's presentation(Tang poetry and Songci generation and Intelligent QA system) for Tsinghua University's 105 anniversary.<br />
* Submit our IJCAI paper to arxiv. (Solve a big problem about submitting the paper including Chinese chacracters. [http://cslt.riit.tsinghua.edu.cn/mediawiki/index.php/How_to_submit_the_latex_files_including_Chinese_characters_to_arxiv Solution])<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
--------------------2016-04-22<br />
* Speed up process of the test performance about theano version of Generationg the similar questions' vectors based on RNN.<br />
--------------------2016-04-23<br />
* Use entity match rule 1(dpk, ix, lpk sv + machine number) to improve the accuracy from 38% to 58%.<br />
--------------------2016-04-24<br />
* Use entity match rules 2+3(dps, s, e, fp, tps, q, 200 + machine number and machine number only) to improve the accuracy from 58% to 72.09%.<br />
* Use entity match rules 4(dpk, ix, lpk sv, dps, s, e, fp, tps, q + machine number + English chacracters) to improve the accuracy from 72.09% to 72.86%.<br />
* Use entity match rules 5(machine number + English chacracters) to improve the accuracy from 72.86% to 76.48%.<br />
* Use simlar pair match rules 1(“快”和“速度”、“复写”和“拷贝”等等) to improve the accuracy from 76.48% to ？%.<br />
=== Plan to do next week ===<br />
* To implement tensorflow version of RNN/LSTM Max margin vector training.<br />
* To implement attention chatting model with xiaobing corpus.<br />
===Interested papers ===<br />
*Cascading Bandits: Learning to Rank in the Cascade Model(ICML 2015) [[http://zheng-wen.com/Cascading_Bandit_Paper.pdf pdf]]</div>Ltyhttp://index.cslt.org/mediawiki/index.php/Tianyi_Luo_2016-04-25Tianyi Luo 2016-04-252016-04-24T15:56:02Z<p>Lty：</p>
<hr />
<div>=== Plan to do this week ===<br />
* To implement tensorflow version of RNN/LSTM Max margin vector training.<br />
=== Work done in this week ===<br />
--------------------2016-04-18<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
* Finish implementing theano version of LSTM Max margin vector training.<br />
--------------------2016-04-19<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
--------------------2016-04-20<br />
* Finish submiting the camera version paper of IJCAI 2016.<br />
* Update the version of Technical Report about Chinese Song Iambics generation.<br />
--------------------2016-04-21<br />
* Finish helping Teacher Wang to prepare for text group's presentation(Tang poetry and Songci generation and Intelligent QA system) for Tsinghua University's 105 anniversary.<br />
* Submit our IJCAI paper to arxiv. (Solve a big problem about submitting the paper including Chinese chacracters. [http://cslt.riit.tsinghua.edu.cn/mediawiki/index.php/How_to_submit_the_latex_files_including_Chinese_characters_to_arxiv Solution])<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
--------------------2016-04-22<br />
* Speed up process of the test performance about theano version of Generationg the similar questions' vectors based on RNN.<br />
--------------------2016-04-23<br />
* Use entity match rule 1(dpk, ix, lpk sv + machine number) to improve the accuracy from 38% to 58%.<br />
--------------------2016-04-24<br />
* Use entity match rules 2+3(dps, s, e, fp, tps, q, 200 + machine number and machine number only) to improve the accuracy from 58% to 72.09%.<br />
* Use entity match rules 4(dpk, ix, lpk sv, dps, s, e, fp, tps, q + machine number + English chacracters) to improve the accuracy from 72.09% to 72.86%.<br />
* Use entity match rules 5(machine number + English chacracters) to improve the accuracy from 72.86% to 76.48%.<br />
* Use simlar pair match rules 1(“快”和“速度”、“复写”和“拷贝”等等) to improve the accuracy from 72.86% to 76.48%.<br />
=== Plan to do next week ===<br />
* To implement tensorflow version of RNN/LSTM Max margin vector training.<br />
* To implement attention chatting model with xiaobing corpus.<br />
===Interested papers ===<br />
*Cascading Bandits: Learning to Rank in the Cascade Model(ICML 2015) [[http://zheng-wen.com/Cascading_Bandit_Paper.pdf pdf]]</div>Ltyhttp://index.cslt.org/mediawiki/index.php/Tianyi_Luo_2016-04-25Tianyi Luo 2016-04-252016-04-24T15:42:24Z<p>Lty：</p>
<hr />
<div>=== Plan to do this week ===<br />
* To implement tensorflow version of RNN/LSTM Max margin vector training.<br />
=== Work done in this week ===<br />
--------------------2016-04-18<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
* Finish implementing theano version of LSTM Max margin vector training.<br />
--------------------2016-04-19<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
--------------------2016-04-20<br />
* Finish submiting the camera version paper of IJCAI 2016.<br />
* Update the version of Technical Report about Chinese Song Iambics generation.<br />
--------------------2016-04-21<br />
* Finish helping Teacher Wang to prepare for text group's presentation(Tang poetry and Songci generation and Intelligent QA system) for Tsinghua University's 105 anniversary.<br />
* Submit our IJCAI paper to arxiv. (Solve a big problem about submitting the paper including Chinese chacracters. [http://cslt.riit.tsinghua.edu.cn/mediawiki/index.php/How_to_submit_the_latex_files_including_Chinese_characters_to_arxiv Solution])<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
--------------------2016-04-22<br />
* Speed up process of the test performance about theano version of Generationg the similar questions' vectors based on RNN.<br />
--------------------2016-04-23<br />
* Use entity match rule 1(dpk, ix, lpk sv + machine number) to improve the accuracy from 38% to 58%.<br />
--------------------2016-04-24<br />
* Use entity match rules 2+3(dps, s, e, fp, tps, q, 200 + machine number and machine number only) to improve the accuracy from 58% to 72.09%.<br />
* Use entity match rules 4(dpk, ix, lpk sv, dps, s, e, fp, tps, q + machine number + English chacracters) to improve the accuracy from 72.09% to 72.86%.<br />
* Use entity match rules 5(machine number + English chacracters) to improve the accuracy from 72.86% to 76.48%.<br />
=== Plan to do next week ===<br />
* To implement tensorflow version of RNN/LSTM Max margin vector training.<br />
* To implement attention chatting model with xiaobing corpus.<br />
===Interested papers ===<br />
*Cascading Bandits: Learning to Rank in the Cascade Model(ICML 2015) [[http://zheng-wen.com/Cascading_Bandit_Paper.pdf pdf]]</div>Ltyhttp://index.cslt.org/mediawiki/index.php/Tianyi_Luo_2016-04-25Tianyi Luo 2016-04-252016-04-24T14:36:44Z<p>Lty：</p>
<hr />
<div>=== Plan to do this week ===<br />
* To implement tensorflow version of RNN/LSTM Max margin vector training.<br />
=== Work done in this week ===<br />
--------------------2016-04-18<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
* Finish implementing theano version of LSTM Max margin vector training.<br />
--------------------2016-04-19<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
--------------------2016-04-20<br />
* Finish submiting the camera version paper of IJCAI 2016.<br />
* Update the version of Technical Report about Chinese Song Iambics generation.<br />
--------------------2016-04-21<br />
* Finish helping Teacher Wang to prepare for text group's presentation(Tang poetry and Songci generation and Intelligent QA system) for Tsinghua University's 105 anniversary.<br />
* Submit our IJCAI paper to arxiv. (Solve a big problem about submitting the paper including Chinese chacracters. [http://cslt.riit.tsinghua.edu.cn/mediawiki/index.php/How_to_submit_the_latex_files_including_Chinese_characters_to_arxiv Solution])<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
--------------------2016-04-22<br />
* Speed up process of the test performance about theano version of Generationg the similar questions' vectors based on RNN.<br />
--------------------2016-04-23<br />
* Use entity match rule 1(dpk, ix, lpk sv + machine number) to improve the accuracy from 38% to 58%.<br />
--------------------2016-04-24<br />
* Use entity match rules 2+3(dps, s, e, fp, tps, q, 200 + machine number and machine number only) to improve the accuracy from 58% to 72.09%.<br />
* Use entity match rules 4(dpk, ix, lpk sv, dps, s, e, fp, tps, q + machine number + English chacracters) to improve the accuracy from 72.09% to 72.86%.<br />
* Use entity match rules 5(machine number + English chacracters) to improve the accuracy from 72.86% to ?%.<br />
=== Plan to do next week ===<br />
* To implement tensorflow version of RNN/LSTM Max margin vector training.<br />
* To implement attention chatting model with xiaobing corpus.<br />
===Interested papers ===<br />
*Cascading Bandits: Learning to Rank in the Cascade Model(ICML 2015) [[http://zheng-wen.com/Cascading_Bandit_Paper.pdf pdf]]</div>Ltyhttp://index.cslt.org/mediawiki/index.php/Tianyi_Luo_2016-04-25Tianyi Luo 2016-04-252016-04-24T14:35:26Z<p>Lty：</p>
<hr />
<div>=== Plan to do this week ===<br />
* To implement tensorflow version of RNN/LSTM Max margin vector training.<br />
=== Work done in this week ===<br />
--------------------2016-04-18<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
* Finish implementing theano version of LSTM Max margin vector training.<br />
--------------------2016-04-19<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
--------------------2016-04-20<br />
* Finish submiting the camera version paper of IJCAI 2016.<br />
* Update the version of Technical Report about Chinese Song Iambics generation.<br />
--------------------2016-04-21<br />
* Finish helping Teacher Wang to prepare for text group's presentation(Tang poetry and Songci generation and Intelligent QA system) for Tsinghua University's 105 anniversary.<br />
* Submit our IJCAI paper to arxiv. (Solve a big problem about submitting the paper including Chinese chacracters. [http://cslt.riit.tsinghua.edu.cn/mediawiki/index.php/How_to_submit_the_latex_files_including_Chinese_characters_to_arxiv Solution])<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
--------------------2016-04-22<br />
* Speed up process of the test performance about theano version of Generationg the similar questions' vectors based on RNN.<br />
--------------------2016-04-23<br />
* Use entity match rule 1(dpk, ix, lpk sv + machine number) to improve the accuracy from 38% to 58%.<br />
--------------------2016-04-24<br />
* Use entity match rules 2+3(dps, s, e, fp, tps, q, 200 + machine number and machine number only) to improve the accuracy from 58% to 72.09%.<br />
* Use entity match rules 4(dpk, ix, lpk sv, dps, s, e, fp, tps, q + machine number + English chacracters) to improve the accuracy from 72.09% to 72.86%.<br />
=== Plan to do next week ===<br />
* To implement tensorflow version of RNN/LSTM Max margin vector training.<br />
* To implement attention chatting model with xiaobing corpus.<br />
===Interested papers ===<br />
*Cascading Bandits: Learning to Rank in the Cascade Model(ICML 2015) [[http://zheng-wen.com/Cascading_Bandit_Paper.pdf pdf]]</div>Ltyhttp://index.cslt.org/mediawiki/index.php/Tianyi_Luo_2016-04-25Tianyi Luo 2016-04-252016-04-24T14:12:04Z<p>Lty：</p>
<hr />
<div>=== Plan to do this week ===<br />
* To implement tensorflow version of RNN/LSTM Max margin vector training.<br />
=== Work done in this week ===<br />
--------------------2016-04-18<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
* Finish implementing theano version of LSTM Max margin vector training.<br />
--------------------2016-04-19<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
--------------------2016-04-20<br />
* Finish submiting the camera version paper of IJCAI 2016.<br />
* Update the version of Technical Report about Chinese Song Iambics generation.<br />
--------------------2016-04-21<br />
* Finish helping Teacher Wang to prepare for text group's presentation(Tang poetry and Songci generation and Intelligent QA system) for Tsinghua University's 105 anniversary.<br />
* Submit our IJCAI paper to arxiv. (Solve a big problem about submitting the paper including Chinese chacracters. [http://cslt.riit.tsinghua.edu.cn/mediawiki/index.php/How_to_submit_the_latex_files_including_Chinese_characters_to_arxiv Solution])<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
--------------------2016-04-22<br />
* Speed up process of the test performance about theano version of Generationg the similar questions' vectors based on RNN.<br />
--------------------2016-04-23<br />
* Use entity match rule 1(dpk, ix, lpk sv + machine number) to improve the accuracy from 38% to 58%.<br />
--------------------2016-04-24<br />
* Use entity match rules 2+3(dps, s, e, fp, tps, q, 200 + machine number and machine number only) to improve the accuracy from 58% to 72%.<br />
* Use entity match rules 4(dpk, ix, lpk sv, dps, s, e, fp, tps, q + machine number + English chacracters) to improve the accuracy from 58% to 72%.<br />
=== Plan to do next week ===<br />
* To implement tensorflow version of RNN/LSTM Max margin vector training.<br />
* To implement attention chatting model with xiaobing corpus.<br />
===Interested papers ===<br />
*Cascading Bandits: Learning to Rank in the Cascade Model(ICML 2015) [[http://zheng-wen.com/Cascading_Bandit_Paper.pdf pdf]]</div>Ltyhttp://index.cslt.org/mediawiki/index.php/Tianyi_Luo_2016-04-25Tianyi Luo 2016-04-252016-04-24T13:36:06Z<p>Lty：</p>
<hr />
<div>=== Plan to do this week ===<br />
* To implement tensorflow version of RNN/LSTM Max margin vector training.<br />
=== Work done in this week ===<br />
--------------------2016-04-18<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
* Finish implementing theano version of LSTM Max margin vector training.<br />
--------------------2016-04-19<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
--------------------2016-04-20<br />
* Finish submiting the camera version paper of IJCAI 2016.<br />
* Update the version of Technical Report about Chinese Song Iambics generation.<br />
--------------------2016-04-21<br />
* Finish helping Teacher Wang to prepare for text group's presentation(Tang poetry and Songci generation and Intelligent QA system) for Tsinghua University's 105 anniversary.<br />
* Submit our IJCAI paper to arxiv. (Solve a big problem about submitting the paper including Chinese chacracters. [http://cslt.riit.tsinghua.edu.cn/mediawiki/index.php/How_to_submit_the_latex_files_including_Chinese_characters_to_arxiv Solution])<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
--------------------2016-04-22<br />
* Speed up process of the test performance about theano version of Generationg the similar questions' vectors based on RNN.<br />
--------------------2016-04-23<br />
* Use entity match rules(dpk, ix, lpk sv + machine number) to improve the accuracy from 38% to 58%.<br />
--------------------2016-04-24<br />
* Use entity match rules(dps, s, e, fp, tps, q, 200 + machine number and machine number only) to improve the accuracy from 58% to 72%.<br />
=== Plan to do next week ===<br />
* To implement tensorflow version of RNN/LSTM Max margin vector training.<br />
* To implement attention chatting model with xiaobing corpus.<br />
===Interested papers ===<br />
*Cascading Bandits: Learning to Rank in the Cascade Model(ICML 2015) [[http://zheng-wen.com/Cascading_Bandit_Paper.pdf pdf]]</div>Ltyhttp://index.cslt.org/mediawiki/index.php/Tianyi_Luo_2016-04-25Tianyi Luo 2016-04-252016-04-24T13:22:05Z<p>Lty：</p>
<hr />
<div>=== Plan to do this week ===<br />
* To implement tensorflow version of RNN/LSTM Max margin vector training.<br />
=== Work done in this week ===<br />
--------------------2016-04-18<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
* Finish implementing theano version of LSTM Max margin vector training.<br />
--------------------2016-04-19<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
--------------------2016-04-20<br />
* Finish submiting the camera version paper of IJCAI 2016.<br />
* Update the version of Technical Report about Chinese Song Iambics generation.<br />
--------------------2016-04-21<br />
* Finish helping Teacher Wang to prepare for text group's presentation(Tang poetry and Songci generation and Intelligent QA system) for Tsinghua University's 105 anniversary.<br />
* Submit our IJCAI paper to arxiv. (Solve a big problem about submitting the paper including Chinese chacracters. [http://cslt.riit.tsinghua.edu.cn/mediawiki/index.php/How_to_submit_the_latex_files_including_Chinese_characters_to_arxiv Solution])<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
--------------------2016-04-22<br />
* Speed up process of the test performance about theano version of Generationg the similar questions' vectors based on RNN.<br />
--------------------2016-04-23<br />
* Use entity match rules to improve the accuracy from 38% to 58%.<br />
--------------------2016-04-24<br />
* Use entity match rules to improve the accuracy from 58% to 72%.<br />
=== Plan to do next week ===<br />
* To implement tensorflow version of RNN/LSTM Max margin vector training.<br />
* To implement attention chatting model with xiaobing corpus.<br />
===Interested papers ===<br />
*Cascading Bandits: Learning to Rank in the Cascade Model(ICML 2015) [[http://zheng-wen.com/Cascading_Bandit_Paper.pdf pdf]]</div>Ltyhttp://index.cslt.org/mediawiki/index.php/Tianyi_Luo_2016-04-25Tianyi Luo 2016-04-252016-04-24T13:21:01Z<p>Lty：</p>
<hr />
<div>=== Plan to do this week ===<br />
* To implement tensorflow version of RNN/LSTM Max margin vector training.<br />
=== Work done in this week ===<br />
--------------------2016-04-18<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
* Finish implementing theano version of LSTM Max margin vector training.<br />
--------------------2016-04-19<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
--------------------2016-04-20<br />
* Finish submiting the camera version paper of IJCAI 2016.<br />
* Update the version of Technical Report about Chinese Song Iambics generation.<br />
--------------------2016-04-21<br />
* Finish helping Teacher Wang to prepare for text group's presentation(Tang poetry and Songci generation and Intelligent QA system) for Tsinghua University's 105 anniversary.<br />
* Submit our IJCAI paper to arxiv. (Solve a big problem about submitting the paper including Chinese chacracters)<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
--------------------2016-04-22<br />
* Speed up process of the test performance about theano version of Generationg the similar questions' vectors based on RNN.<br />
--------------------2016-04-23<br />
* Use entity match rules to improve the accuracy from 38% to 58%.<br />
--------------------2016-04-24<br />
* Use entity match rules to improve the accuracy from 58% to 72%.<br />
=== Plan to do next week ===<br />
* To implement tensorflow version of RNN/LSTM Max margin vector training.<br />
* To implement attention chatting model with xiaobing corpus.<br />
===Interested papers ===<br />
*Cascading Bandits: Learning to Rank in the Cascade Model(ICML 2015) [[http://zheng-wen.com/Cascading_Bandit_Paper.pdf pdf]]</div>Ltyhttp://index.cslt.org/mediawiki/index.php/Tianyi_Luo_2016-04-25Tianyi Luo 2016-04-252016-04-24T13:20:05Z<p>Lty：</p>
<hr />
<div>=== Plan to do this week ===<br />
* To implement tensorflow version of RNN/LSTM Max margin vector training.<br />
=== Work done in this week ===<br />
--------------------2016-04-18<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
* Finish implementing theano version of LSTM Max margin vector training.<br />
--------------------2016-04-19<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
--------------------2016-04-20<br />
* Finish submiting the camera version paper of IJCAI 2016.<br />
* Update the version of Technical Report about Chinese Song Iambics generation.<br />
--------------------2016-04-21<br />
* Finish helping Teacher Wang to prepare for text group's presentation(Tang poetry and Songci generation and Intelligent QA system) for Tsinghua University's 105 anniversary.<br />
* Submit our IJCAI paper to arxiv. (Solve a big problem about submitting the paper including Chinese chacracters)<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
--------------------2016-04-22<br />
* Speed up process of the test performance about theano version of Generationg the similar questions' vectors based on RNN.<br />
--------------------2016-04-23<br />
* Use entity match rules to improve the accuracy from 38% to 58%.<br />
--------------------2016-04-24<br />
* Use entity match rules to improve the accuracy from 58% to 72%.<br />
=== Plan to do next week ===<br />
* To implement tensorflow version of RNN/LSTM Max margin vector training.<br />
===Interested papers ===<br />
*Cascading Bandits: Learning to Rank in the Cascade Model(ICML 2015) [[http://zheng-wen.com/Cascading_Bandit_Paper.pdf pdf]]</div>Ltyhttp://index.cslt.org/mediawiki/index.php/ScheduleSchedule2016-04-23T10:42:40Z<p>Lty：</p>
<hr />
<div>=Text Processing Team Schedule=<br />
<br />
==Members==<br />
===Former Members===<br />
* Rong Liu (刘荣) : 优酷<br />
* Xiaoxi Wang (王晓曦) : 图灵机器人<br />
* Xi Ma (马习) : 清华大学研究生<br />
* DongXu Zhang (张东旭) : --<br />
<br />
===Current Members===<br />
* Tianyi Luo (骆天一)<br />
* Chao Xing (邢超)<br />
* Qixin Wang (王琪鑫)<br />
* Yiqiao Pan (潘一桥)<br />
<br />
==Work Process==<br />
===Similar questions senetence vector model training with RNN/LSTM and the attention RNN/LSTM chatting model training (Tianyi Luo)===<br />
--------------------2016-04-22<br />
* Speed up process of the test performance about theano version of Generationg the similar questions' vectors based on RNN.<br />
--------------------2016-04-21<br />
* Finish helping Teacher Wang to prepare for text group's presentation(Tang poetry and Songci generation and Intelligent QA system) for Tsinghua University's 105 anniversary.<br />
* Submit our IJCAI paper to arxiv. (Solve a big problem about submitting the paper including Chinese chacracters. [http://cslt.riit.tsinghua.edu.cn/mediawiki/index.php/How_to_submit_the_latex_files_including_Chinese_characters_to_arxiv Solution])<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
--------------------2016-04-20<br />
* Finish submiting the camera version paper of IJCAI 2016.<br />
* Update the version of Technical Report about Chinese Song Iambics generation.<br />
--------------------2016-04-19<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
--------------------2016-04-18<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
* Finish implementing theano version of LSTM Max margin vector training.<br />
===Reproduce DSSM Baseline (Chao Xing)===<br />
: 2016-04-23 : Set a series of experiment set.<br />
1. Try deep CNN-DSSM, current model just follow proposed model contain one convolution layer, need to be a tuneable parameter.<br />
2. Test whether mixture data effective to current model and deep CDSSM.<br />
3. Code Recurrent CNN-DSSM (new approach.)<br />
: 2016-04-22 : Find a problem : Use labs' gpu machine 970 iteration per time is 1537 second but huilan's server is just 7 second.<br />
Achieve reasonable results when apply max-margin method to CNN-DSSM model.<br />
: 2016-04-21 : True DSSM model doesn't work well, analysis as below:<br />
1. Not exactly reproduce DSSM model, because the original one is English version, I just adapt it to Chinese but after word segmentation. <br />
So the input is tri-gram words not tri-gram letter.<br />
2. Our dataset far from rich, because of we do not use pre-trained word vectors as initial vectors, we can hardly achieve good performance.<br />
: Request<br />
1. As we have rich pre-trained word vectors, maybe CDSSM or RDSSM corrected to our task.<br />
2. Different length of sequences seek to be fixed dimension vectors, just CNN and RNN can do such things, DNN can not do it by using <br />
fix length of word vectors<br />
: Coding done CDSSM. Test for it's performance.<br />
One problem : When you install tensorflow by pip 0.8.0 and you want to use conv2d function by gpu, you need make sure you had already <br />
install your cudnn's version as 4.0 not lastest 5.0.<br />
: 2016-04-20 : Find reproduced DSSM model's bug, fix it.<br />
: 2016-04-19 : Code mixture data model by less memory dependency done. Test it's performance.<br />
: 2016-04-18 : Code mixture data model.<br />
: 2016-04-16 : Code mixture data model, but face to memory error. Dr. Wang help me fix it.<br />
: 2016-04-15 : Share Papers. Investigation a series of DSSM papers for future work. And show our intern students how to do research.<br />
: Original DSSM model : Learning Deep Structured Semantic Models for Web Search using Clickthrough Data [http://cslt.riit.tsinghua.edu.cn/mediawiki/images/4/45/2013_-_Learning_Deep_Structured_Semantic_Models_for_Web_Search_using_Clickthrough_Data_-_Report.pdf pdf]<br />
: CNN based DSSM model : A Latent Semantic Model with Convolutional-Pooling Structure for Information Retrieval [http://cslt.riit.tsinghua.edu.cn/mediawiki/images/b/b7/2014_-_A_Latent_Semantic_Model_with_Convolutional-Pooling_Structure_for_Information_Retrieval_-_Report.pdf pdf]<br />
: Use DSSM model for a new area : Modeling Interestingness with Deep Neural Networks [http://cslt.riit.tsinghua.edu.cn/mediawiki/images/1/1f/2014_-_Modeling_Interestingness_with_Deep_Neural_Networks_-_Report.pdf pdf]<br />
: Latest approach for LSTM + RNN DSSM model : SEMANTIC MODELLING WITH LONG-SHORT-TERM MEMORY FOR INFORMATION RETRIEVAL [http://cslt.riit.tsinghua.edu.cn/mediawiki/images/2/24/2015_-_SEMANTIC_MODELLING_WITH_LONG-SHORT-TERM_MEMORY_FOR_INFORMATION_RETRIEVAL_-_Report.pdf pdf]<br />
<br />
: 2016-04-14 : Test dssm-dnn model, code dssm-cnn model.<br />
Continue investigate deep neural question answering system.<br />
: 2016-04-13 : test dssm model, investigate deep neural question answering system.<br />
: Share theano ppt [http://cslt.riit.tsinghua.edu.cn/mediawiki/index.php/%E6%96%87%E4%BB%B6:Theano-RBM.pptx theano]<br />
: Share tensorflow ppt [http://cslt.riit.tsinghua.edu.cn/mediawiki/index.php/%E6%96%87%E4%BB%B6:Tensorflow.pptx tensorflow]<br />
: 2016-04-12 : Write done dssm tensor flow version.<br />
: 2016-04-11 : Write tensorflow toolkit ppt for intern student.<br />
: 2016-04-10 : Learn tensorflow toolkit.<br />
: 2016-04-09 : Learn tensorflow toolkit.<br />
: 2016-04-08 : Finish theano version.<br />
<br />
===Deep Poem Processing With Image (Ziwei Bai)===<br />
: 2016-04-20 :combine my program with Qixin Wang's<br />
: 2016-04-10 : web spider to catch a thousand pices of images.<br />
: 2016-04-13 :1、download theano for python2.7。 2.debug cnn.py<br />
: 2016-04-15 :web spider to catch 30 thousands pices of images and store them into a matrix<br />
: 2016-04-16 :modify the code of CNN and spider<br />
: 2016-04-17 :train convouloutional neural network<br />
<br />
===RNN Music Processing for lyric (Shiyao Li)===<br />
: 2016-04-20 : learn LSTM<br />
: 2016-04-09 : web spider to catch a thousand pieces of lyrics.<br />
: 2016-04-10 : extract the keywords in the lyrics<br />
: 2016-04-13 :Read paper Memory Network.<br />
: 2016-04-15 :read the paper Memory Network and start to understand its code<br />
: 2016-04-17 :read paper end to end memory network<br />
<br />
===RNN Key word Poem Processing (Yi Xiong)===<br />
: 2016-04-22 : Code a web spider to recursively catch link of keywords from Baidu<br />
: 2016-04-09 : Database for N-Gram data storing<br />
: 2016-04-10 : dictionary stored in database , dictionary based segmentation and a simple bigram segmentation<br />
: 2016-04-13 : segmentation result analysis<br />
: 2016-04-15 :improve the simple bigram segmentation<br />
: 2016-04-16 :compare the result of bigram segmentation with dictionary segmentation<br />
: 2016-04-17 :learn python (head first 50%)<br />
: 2016-04-20 : learn web spider<br />
<br />
===RNN Piano Processing (Jiyuan Zhang)===<br />
:2016-4-12：select appropriate midis and run rnnrbm model<br />
:2016-4-13：view rnnrbm model‘s code<br />
:2016-4-14~15:coding to select 4/4 beat of midis<br />
:2016-4-17~22:run data, failed several times ，then modify code and view rnnrbm model's code<br />
<br />
===Recommendation System (Tong Liu)===<br />
: 2016-04-09 : 1.read a review:Machine learning:Trends,perspectives, and prospects 2.learn python ,can operate dict and set<br />
: 2016-04-12 : 1.read paper Collaborative Deep Learning for Recommender Systems and take notes.2. learn the concepts of stacked denoising autoencoder(SDAE).<br />
: 2016-04-17 :1.allocate PuTTy and Xming 2.learn python, can operate slice and iterator 3.learn release and datasets of a paper: Collaborative Deep Learning for Recommender Systems<br />
<br />
===Question & Answering (Aiting Liu)===<br />
: 2016-04-20 : read Fader's paper ()2013 <br />
: 2016-04-15 :learn dssm and sent2vec<br />
: 2016-04-16 :try to figure out how thePARALAX dataset is constructed<br />
: 2016-04-17 :download the PARALAX dataset and turn it into what we want it to be</div>Ltyhttp://index.cslt.org/mediawiki/index.php/ScheduleSchedule2016-04-22T03:25:36Z<p>Lty：</p>
<hr />
<div>=Text Processing Team Schedule=<br />
<br />
==Members==<br />
===Former Members===<br />
* Rong Liu (刘荣) : 优酷<br />
* Xiaoxi Wang (王晓曦) : 图灵机器人<br />
* Xi Ma (马习) : 清华大学研究生<br />
* DongXu Zhang (张东旭) : --<br />
<br />
===Current Members===<br />
* Tianyi Luo (骆天一)<br />
* Chao Xing (邢超)<br />
* Qixin Wang (王琪鑫)<br />
* Yiqiao Pan (潘一桥)<br />
<br />
==Work Process==<br />
===Similar questions senetence vector model training with RNN/LSTM and the attention RNN/LSTM chatting model training (Tianyi Luo)===<br />
--------------------2016-04-21<br />
* Finish helping Teacher Wang to prepare for text group's presentation(Tang poetry and Songci generation and Intelligent QA system) for Tsinghua University's 105 anniversary.<br />
* Submit our IJCAI paper to arxiv. (Solve a big problem about submitting the paper including Chinese chacracters. [http://cslt.riit.tsinghua.edu.cn/mediawiki/index.php/How_to_submit_the_latex_files_including_Chinese_characters_to_arxiv Solution])<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
--------------------2016-04-20<br />
* Finish submiting the camera version paper of IJCAI 2016.<br />
* Update the version of Technical Report about Chinese Song Iambics generation.<br />
--------------------2016-04-19<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
--------------------2016-04-18<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
* Finish implementing theano version of LSTM Max margin vector training.<br />
===Reproduce DSSM Baseline (Chao Xing)===<br />
: 2016-04-21 : True DSSM model doesn't work well, analysis as below:<br />
1. Not exactly reproduce DSSM model, because the original one is English version, I just adapt it to Chinese but after word segmentation. <br />
So the input is tri-gram words not tri-gram letter.<br />
2. Our dataset far from rich, because of we do not use pre-trained word vectors as initial vectors, we can hardly achieve good performance.<br />
: Request<br />
1. As we have rich pre-trained word vectors, maybe CDSSM or RDSSM corrected to our task.<br />
2. Different length of sequences seek to be fixed dimension vectors, just CNN and RNN can do such things, DNN can not do it by using <br />
fix length of word vectors<br />
: Coding done CDSSM. Test for it's performance.<br />
One problem : When you install tensorflow by pip 0.8.0 and you want to use conv2d function by gpu, you need make sure you had already <br />
install your cudnn's version as 4.0 not lastest 5.0.<br />
: 2016-04-20 : Find reproduced DSSM model's bug, fix it.<br />
: 2016-04-19 : Code mixture data model by less memory dependency done. Test it's performance.<br />
: 2016-04-18 : Code mixture data model.<br />
: 2016-04-16 : Code mixture data model, but face to memory error. Dr. Wang help me fix it.<br />
: 2016-04-15 : Share Papers. Investigation a series of DSSM papers for future work. And show our intern students how to do research.<br />
: Original DSSM model : Learning Deep Structured Semantic Models for Web Search using Clickthrough Data [http://cslt.riit.tsinghua.edu.cn/mediawiki/images/4/45/2013_-_Learning_Deep_Structured_Semantic_Models_for_Web_Search_using_Clickthrough_Data_-_Report.pdf pdf]<br />
: CNN based DSSM model : A Latent Semantic Model with Convolutional-Pooling Structure for Information Retrieval [http://cslt.riit.tsinghua.edu.cn/mediawiki/images/b/b7/2014_-_A_Latent_Semantic_Model_with_Convolutional-Pooling_Structure_for_Information_Retrieval_-_Report.pdf pdf]<br />
: Use DSSM model for a new area : Modeling Interestingness with Deep Neural Networks [http://cslt.riit.tsinghua.edu.cn/mediawiki/images/1/1f/2014_-_Modeling_Interestingness_with_Deep_Neural_Networks_-_Report.pdf pdf]<br />
: Latest approach for LSTM + RNN DSSM model : SEMANTIC MODELLING WITH LONG-SHORT-TERM MEMORY FOR INFORMATION RETRIEVAL [http://cslt.riit.tsinghua.edu.cn/mediawiki/images/2/24/2015_-_SEMANTIC_MODELLING_WITH_LONG-SHORT-TERM_MEMORY_FOR_INFORMATION_RETRIEVAL_-_Report.pdf pdf]<br />
<br />
: 2016-04-14 : Test dssm-dnn model, code dssm-cnn model.<br />
Continue investigate deep neural question answering system.<br />
: 2016-04-13 : test dssm model, investigate deep neural question answering system.<br />
: Share theano ppt [http://cslt.riit.tsinghua.edu.cn/mediawiki/index.php/%E6%96%87%E4%BB%B6:Theano-RBM.pptx theano]<br />
: Share tensorflow ppt [http://cslt.riit.tsinghua.edu.cn/mediawiki/index.php/%E6%96%87%E4%BB%B6:Tensorflow.pptx tensorflow]<br />
: 2016-04-12 : Write done dssm tensor flow version.<br />
: 2016-04-11 : Write tensorflow toolkit ppt for intern student.<br />
: 2016-04-10 : Learn tensorflow toolkit.<br />
: 2016-04-09 : Learn tensorflow toolkit.<br />
: 2016-04-08 : Finish theano version.<br />
<br />
===Deep Poem Processing With Image (Ziwei Bai)===<br />
: 2016-04-20 :combine my program with Qixin Wang's<br />
: 2016-04-10 : web spider to catch a thousand pices of images.<br />
: 2016-04-13 :1、download theano for python2.7。 2.debug cnn.py<br />
: 2016-04-15 :web spider to catch 30 thousands pices of images and store them into a matrix<br />
: 2016-04-16 :modify the code of CNN and spider<br />
: 2016-04-17 :train convouloutional neural network<br />
<br />
===RNN Music Processing for lyric (Shiyao Li)===<br />
: 2016-04-20 : learn LSTM<br />
: 2016-04-09 : web spider to catch a thousand pieces of lyrics.<br />
: 2016-04-10 : extract the keywords in the lyrics<br />
: 2016-04-13 :Read paper Memory Network.<br />
: 2016-04-15 :read the paper Memory Network and start to understand its code<br />
: 2016-04-17 :read paper end to end memory network<br />
<br />
===RNN Key word Poem Processing (Yi Xiong)===<br />
: 2016-04-20 : learn web spider<br />
: 2016-04-09 : Database for N-Gram data storing<br />
: 2016-04-10 : dictionary stored in database , dictionary based segmentation and a simple bigram segmentation<br />
: 2016-04-13 : segmentation result analysis<br />
: 2016-04-15 :improve the simple bigram segmentation<br />
: 2016-04-16 :compare the result of bigram segmentation with dictionary segmentation<br />
: 2016-04-17 :learn python (head first 50%)<br />
<br />
===RNN Piano Processing (Jiyuan Zhang)===<br />
:2016-4-12：select appropriate midis and run rnnrbm model<br />
:2016-4-13：view rnnrbm model‘s code<br />
<br />
===Recommendation System (Tong Liu)===<br />
: 2016-04-09 : 1.read a review:Machine learning:Trends,perspectives, and prospects 2.learn python ,can operate dict and set<br />
: 2016-04-12 : 1.read paper Collaborative Deep Learning for Recommender Systems and take notes.2. learn the concepts of stacked denoising autoencoder(SDAE).<br />
: 2016-04-17 :1.allocate PuTTy and Xming 2.learn python, can operate slice and iterator 3.learn release and datasets of a paper: Collaborative Deep Learning for Recommender Systems<br />
<br />
===Question & Answering (Aiting Liu)===<br />
: 2016-04-20 : read Fader's paper ()2013 <br />
: 2016-04-15 :learn dssm and sent2vec<br />
: 2016-04-16 :try to figure out how thePARALAX dataset is constructed<br />
: 2016-04-17 :download the PARALAX dataset and turn it into what we want it to be</div>Ltyhttp://index.cslt.org/mediawiki/index.php/ScheduleSchedule2016-04-22T03:10:42Z<p>Lty：</p>
<hr />
<div>=Text Processing Team Schedule=<br />
<br />
==Members==<br />
===Former Members===<br />
* Rong Liu (刘荣) : 优酷<br />
* Xiaoxi Wang (王晓曦) : 图灵机器人<br />
* Xi Ma (马习) : 清华大学研究生<br />
* DongXu Zhang (张东旭) : --<br />
<br />
===Current Members===<br />
* Tianyi Luo (骆天一)<br />
* Chao Xing (邢超)<br />
* Qixin Wang (王琪鑫)<br />
* Yiqiao Pan (潘一桥)<br />
<br />
==Work Process==<br />
===Similar questions senetence vector model training with RNN/LSTM and the attention RNN/LSTM chatting model training (Tianyi Luo)===<br />
--------------------2016-04-18<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
* Finish implementing theano version of LSTM Max margin vector training.<br />
--------------------2016-04-19<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
--------------------2016-04-20<br />
* Finish submiting the camera version paper of IJCAI 2016.<br />
* Update the version of Technical Report about Chinese Song Iambics generation.<br />
--------------------2016-04-21<br />
* Finish helping Teacher Wang to prepare for text group's presentation(Tang poetry and Songci generation and Intelligent QA system) for Tsinghua University's 105 anniversary.<br />
* Submit our IJCAI paper to arxiv. (Solve a big problem about submitting the paper including Chinese chacracters. [http://cslt.riit.tsinghua.edu.cn/mediawiki/index.php/How_to_submit_the_latex_files_including_Chinese_characters_to_arxiv Solution])<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
<br />
===Reproduce DSSM Baseline (Chao Xing)===<br />
: 2016-04-21 : True DSSM model doesn't work well, analysis as below:<br />
1. Not exactly reproduce DSSM model, because the original one is English version, I just adapt it to Chinese but after word segmentation. <br />
So the input is tri-gram words not tri-gram letter.<br />
2. Our dataset far from rich, because of we do not use pre-trained word vectors as initial vectors, we can hardly achieve good performance.<br />
: Request :<br />
1. As we have rich pre-trained word vectors, maybe CDSSM or RDSSM corrected to our task.<br />
2. Different length of sequences seek to be fixed dimension vectors, just CNN and RNN can do such things, DNN can not do it by using <br />
fix length of word vectors<br />
: Coding done CDSSM. Test for it's performance.<br />
One problem : When you install tensorflow by pip 0.8.0 and you want to use conv2d function by gpu, you need make sure you had already <br />
install your cudnn's version as 4.0 not lastest 5.0.<br />
: 2016-04-20 : Find reproduced DSSM model's bug, fix it.<br />
: 2016-04-19 : Code mixture data model by less memory dependency done. Test it's performance.<br />
: 2016-04-18 : Code mixture data model.<br />
: 2016-04-16 : Code mixture data model, but face to memory error. Dr. Wang help me fix it.<br />
: 2016-04-15 : Share Papers. Investigation a series of DSSM papers for future work. And show our intern students how to do research.<br />
: Original DSSM model : Learning Deep Structured Semantic Models for Web Search using Clickthrough Data [http://cslt.riit.tsinghua.edu.cn/mediawiki/images/4/45/2013_-_Learning_Deep_Structured_Semantic_Models_for_Web_Search_using_Clickthrough_Data_-_Report.pdf pdf]<br />
: CNN based DSSM model : A Latent Semantic Model with Convolutional-Pooling Structure for Information Retrieval [http://cslt.riit.tsinghua.edu.cn/mediawiki/images/b/b7/2014_-_A_Latent_Semantic_Model_with_Convolutional-Pooling_Structure_for_Information_Retrieval_-_Report.pdf pdf]<br />
: Use DSSM model for a new area : Modeling Interestingness with Deep Neural Networks [http://cslt.riit.tsinghua.edu.cn/mediawiki/images/1/1f/2014_-_Modeling_Interestingness_with_Deep_Neural_Networks_-_Report.pdf pdf]<br />
: Latest approach for LSTM + RNN DSSM model : SEMANTIC MODELLING WITH LONG-SHORT-TERM MEMORY FOR INFORMATION RETRIEVAL [http://cslt.riit.tsinghua.edu.cn/mediawiki/images/2/24/2015_-_SEMANTIC_MODELLING_WITH_LONG-SHORT-TERM_MEMORY_FOR_INFORMATION_RETRIEVAL_-_Report.pdf pdf]<br />
<br />
: 2016-04-14 : Test dssm-dnn model, code dssm-cnn model.<br />
Continue investigate deep neural question answering system.<br />
: 2016-04-13 : test dssm model, investigate deep neural question answering system.<br />
: Share theano ppt [http://cslt.riit.tsinghua.edu.cn/mediawiki/index.php/%E6%96%87%E4%BB%B6:Theano-RBM.pptx theano]<br />
: Share tensorflow ppt [http://cslt.riit.tsinghua.edu.cn/mediawiki/index.php/%E6%96%87%E4%BB%B6:Tensorflow.pptx tensorflow]<br />
: 2016-04-12 : Write done dssm tensor flow version.<br />
: 2016-04-11 : Write tensorflow toolkit ppt for intern student.<br />
: 2016-04-10 : Learn tensorflow toolkit.<br />
: 2016-04-09 : Learn tensorflow toolkit.<br />
: 2016-04-08 : Finish theano version.<br />
<br />
===Deep Poem Processing With Image (Ziwei Bai)===<br />
: 2016-04-20 :combine my program with Qixin Wang's<br />
: 2016-04-10 : web spider to catch a thousand pices of images.<br />
: 2016-04-13 :1、download theano for python2.7。 2.debug cnn.py<br />
: 2016-04-15 :web spider to catch 30 thousands pices of images and store them into a matrix<br />
: 2016-04-16 :modify the code of CNN and spider<br />
: 2016-04-17 :train convouloutional neural network<br />
<br />
===RNN Music Processing for lyric (Shiyao Li)===<br />
: 2016-04-20 : learn LSTM<br />
: 2016-04-09 : web spider to catch a thousand pieces of lyrics.<br />
: 2016-04-10 : extract the keywords in the lyrics<br />
: 2016-04-13 :Read paper Memory Network.<br />
: 2016-04-15 :read the paper Memory Network and start to understand its code<br />
: 2016-04-17 :read paper end to end memory network<br />
<br />
===RNN Key word Poem Processing (Yi Xiong)===<br />
: 2016-04-20 : learn web spider<br />
: 2016-04-09 : Database for N-Gram data storing<br />
: 2016-04-10 : dictionary stored in database , dictionary based segmentation and a simple bigram segmentation<br />
: 2016-04-13 : segmentation result analysis<br />
: 2016-04-15 :improve the simple bigram segmentation<br />
: 2016-04-16 :compare the result of bigram segmentation with dictionary segmentation<br />
: 2016-04-17 :learn python (head first 50%)<br />
<br />
===RNN Piano Processing (Jiyuan Zhang)===<br />
:2016-4-12：select appropriate midis and run rnnrbm model<br />
:2016-4-13：view rnnrbm model‘s code<br />
<br />
===Recommendation System (Tong Liu)===<br />
: 2016-04-09 : 1.read a review:Machine learning:Trends,perspectives, and prospects 2.learn python ,can operate dict and set<br />
: 2016-04-12 : 1.read paper Collaborative Deep Learning for Recommender Systems and take notes.2. learn the concepts of stacked denoising autoencoder(SDAE).<br />
: 2016-04-17 :1.allocate PuTTy and Xming 2.learn python, can operate slice and iterator 3.learn release and datasets of a paper: Collaborative Deep Learning for Recommender Systems<br />
<br />
===Question & Answering (Aiting Liu)===<br />
: 2016-04-20 : read Fader's paper ()2013 <br />
: 2016-04-15 :learn dssm and sent2vec<br />
: 2016-04-16 :try to figure out how thePARALAX dataset is constructed<br />
: 2016-04-17 :download the PARALAX dataset and turn it into what we want it to be</div>Ltyhttp://index.cslt.org/mediawiki/index.php/ScheduleSchedule2016-04-22T03:10:14Z<p>Lty：</p>
<hr />
<div>=Text Processing Team Schedule=<br />
<br />
==Members==<br />
===Former Members===<br />
* Rong Liu (刘荣) : 优酷<br />
* Xiaoxi Wang (王晓曦) : 图灵机器人<br />
* Xi Ma (马习) : 清华大学研究生<br />
* DongXu Zhang (张东旭) : --<br />
<br />
===Current Members===<br />
* Tianyi Luo (骆天一)<br />
* Chao Xing (邢超)<br />
* Qixin Wang (王琪鑫)<br />
* Yiqiao Pan (潘一桥)<br />
<br />
==Work Process==<br />
===Similar questions senetence vector model training with RNN/LSTM and the attention RNN/LSTM chatting model training (Tianyi Luo)===<br />
--------------------2016-04-18<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
* Finish implementing theano version of LSTM Max margin vector training.<br />
--------------------2016-04-19<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
--------------------2016-04-20<br />
* Finish submiting the camera version paper of IJCAI 2016.<br />
* Update the version of Technical Report about Chinese Song Iambics generation.<br />
--------------------2016-04-21<br />
* Finish helping Teacher Wang to prepare for text group's presentation(Tang poetry and Songci generation and Intelligent QA system) for Tsinghua University's 105 anniversary.<br />
* Submit our IJCAI paper to arxiv. (Solve a big problem about submitting the paper including Chinese chacracters. [http://cslt.riit.tsinghua.edu.cn/mediawiki/index.php/How_to_submit_the_latex_files_including_Chinese_characters_to_arxiv Solution]<br />
)<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
<br />
===Reproduce DSSM Baseline (Chao Xing)===<br />
: 2016-04-20 : Find reproduced DSSM model's bug, fix it.<br />
: 2016-04-19 : Code mixture data model by less memory dependency done. Test it's performance.<br />
: 2016-04-18 : Code mixture data model.<br />
: 2016-04-16 : Code mixture data model, but face to memory error. Dr. Wang help me fix it.<br />
: 2016-04-15 : Share Papers. Investigation a series of DSSM papers for future work. And show our intern students how to do research.<br />
: Original DSSM model : Learning Deep Structured Semantic Models for Web Search using Clickthrough Data [http://cslt.riit.tsinghua.edu.cn/mediawiki/images/4/45/2013_-_Learning_Deep_Structured_Semantic_Models_for_Web_Search_using_Clickthrough_Data_-_Report.pdf pdf]<br />
: CNN based DSSM model : A Latent Semantic Model with Convolutional-Pooling Structure for Information Retrieval [http://cslt.riit.tsinghua.edu.cn/mediawiki/images/b/b7/2014_-_A_Latent_Semantic_Model_with_Convolutional-Pooling_Structure_for_Information_Retrieval_-_Report.pdf pdf]<br />
: Use DSSM model for a new area : Modeling Interestingness with Deep Neural Networks [http://cslt.riit.tsinghua.edu.cn/mediawiki/images/1/1f/2014_-_Modeling_Interestingness_with_Deep_Neural_Networks_-_Report.pdf pdf]<br />
: Latest approach for LSTM + RNN DSSM model : SEMANTIC MODELLING WITH LONG-SHORT-TERM MEMORY FOR INFORMATION RETRIEVAL [http://cslt.riit.tsinghua.edu.cn/mediawiki/images/2/24/2015_-_SEMANTIC_MODELLING_WITH_LONG-SHORT-TERM_MEMORY_FOR_INFORMATION_RETRIEVAL_-_Report.pdf pdf]<br />
<br />
: 2016-04-14 : Test dssm-dnn model, code dssm-cnn model.<br />
Continue investigate deep neural question answering system.<br />
: 2016-04-13 : test dssm model, investigate deep neural question answering system.<br />
: Share theano ppt [http://cslt.riit.tsinghua.edu.cn/mediawiki/index.php/%E6%96%87%E4%BB%B6:Theano-RBM.pptx theano]<br />
: Share tensorflow ppt [http://cslt.riit.tsinghua.edu.cn/mediawiki/index.php/%E6%96%87%E4%BB%B6:Tensorflow.pptx tensorflow]<br />
: 2016-04-12 : Write done dssm tensor flow version.<br />
: 2016-04-11 : Write tensorflow toolkit ppt for intern student.<br />
: 2016-04-10 : Learn tensorflow toolkit.<br />
: 2016-04-09 : Learn tensorflow toolkit.<br />
: 2016-04-08 : Finish theano version.<br />
<br />
===Deep Poem Processing With Image (Ziwei Bai)===<br />
: 2016-04-20 :combine my program with Qixin Wang's<br />
: 2016-04-10 : web spider to catch a thousand pices of images.<br />
: 2016-04-13 :1、download theano for python2.7。 2.debug cnn.py<br />
: 2016-04-15 :web spider to catch 30 thousands pices of images and store them into a matrix<br />
: 2016-04-16 :modify the code of CNN and spider<br />
: 2016-04-17 :train convouloutional neural network<br />
<br />
===RNN Music Processing for lyric (Shiyao Li)===<br />
: 2016-04-20 : learn LSTM<br />
: 2016-04-09 : web spider to catch a thousand pieces of lyrics.<br />
: 2016-04-10 : extract the keywords in the lyrics<br />
: 2016-04-13 :Read paper Memory Network.<br />
: 2016-04-15 :read the paper Memory Network and start to understand its code<br />
: 2016-04-17 :read paper end to end memory network<br />
<br />
===RNN Key word Poem Processing (Yi Xiong)===<br />
: 2016-04-20 : learn web spider<br />
: 2016-04-09 : Database for N-Gram data storing<br />
: 2016-04-10 : dictionary stored in database , dictionary based segmentation and a simple bigram segmentation<br />
: 2016-04-13 : segmentation result analysis<br />
: 2016-04-15 :improve the simple bigram segmentation<br />
: 2016-04-16 :compare the result of bigram segmentation with dictionary segmentation<br />
: 2016-04-17 :learn python (head first 50%)<br />
<br />
===RNN Piano Processing (Jiyuan Zhang)===<br />
:2016-4-12：select appropriate midis and run rnnrbm model<br />
:2016-4-13：view rnnrbm model‘s code<br />
<br />
===Recommendation System (Tong Liu)===<br />
: 2016-04-09 : 1.read a review:Machine learning:Trends,perspectives, and prospects 2.learn python ,can operate dict and set<br />
: 2016-04-12 : 1.read paper Collaborative Deep Learning for Recommender Systems and take notes.2. learn the concepts of stacked denoising autoencoder(SDAE).<br />
: 2016-04-17 :1.allocate PuTTy and Xming 2.learn python, can operate slice and iterator 3.learn release and datasets of a paper: Collaborative Deep Learning for Recommender Systems<br />
<br />
===Question & Answering (Aiting Liu)===<br />
: 2016-04-20 : read Fader's paper ()2013 <br />
: 2016-04-15 :learn dssm and sent2vec<br />
: 2016-04-16 :try to figure out how thePARALAX dataset is constructed<br />
: 2016-04-17 :download the PARALAX dataset and turn it into what we want it to be</div>Ltyhttp://index.cslt.org/mediawiki/index.php/ScheduleSchedule2016-04-22T03:06:48Z<p>Lty：</p>
<hr />
<div>=Text Processing Team Schedule=<br />
<br />
==Members==<br />
===Former Members===<br />
* Rong Liu (刘荣) : 优酷<br />
* Xiaoxi Wang (王晓曦) : 图灵机器人<br />
* Xi Ma (马习) : 清华大学研究生<br />
* DongXu Zhang (张东旭) : --<br />
<br />
===Current Members===<br />
* Tianyi Luo (骆天一)<br />
* Chao Xing (邢超)<br />
* Qixin Wang (王琪鑫)<br />
* Yiqiao Pan (潘一桥)<br />
<br />
==Work Process==<br />
===Similar questions senetence vector model training with RNN/LSTM and the attention RNN/LSTM chatting model training (Tianyi Luo)===<br />
--------------------2016-04-18<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
* Finish implementing theano version of LSTM Max margin vector training.<br />
--------------------2016-04-19<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
--------------------2016-04-20<br />
* Finish submiting the camera version paper of IJCAI 2016.<br />
* Update the version of Technical Report about Chinese Song Iambics generation.<br />
--------------------2016-04-21<br />
* Finish helping Teacher Wang to prepare for text group's presentation(Tang poetry and Songci generation and Intelligent QA system) for Tsinghua University's 105 anniversary.<br />
* Submit our IJCAI paper to arxiv. (Solve a big problem about submitting the paper including Chinese chacracters)<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
<br />
===Reproduce DSSM Baseline (Chao Xing)===<br />
: 2016-04-20 : Find reproduced DSSM model's bug, fix it.<br />
: 2016-04-19 : Code mixture data model by less memory dependency done. Test it's performance.<br />
: 2016-04-18 : Code mixture data model.<br />
: 2016-04-16 : Code mixture data model, but face to memory error. Dr. Wang help me fix it.<br />
: 2016-04-15 : Share Papers. Investigation a series of DSSM papers for future work. And show our intern students how to do research.<br />
: Original DSSM model : Learning Deep Structured Semantic Models for Web Search using Clickthrough Data [http://cslt.riit.tsinghua.edu.cn/mediawiki/images/4/45/2013_-_Learning_Deep_Structured_Semantic_Models_for_Web_Search_using_Clickthrough_Data_-_Report.pdf pdf]<br />
: CNN based DSSM model : A Latent Semantic Model with Convolutional-Pooling Structure for Information Retrieval [http://cslt.riit.tsinghua.edu.cn/mediawiki/images/b/b7/2014_-_A_Latent_Semantic_Model_with_Convolutional-Pooling_Structure_for_Information_Retrieval_-_Report.pdf pdf]<br />
: Use DSSM model for a new area : Modeling Interestingness with Deep Neural Networks [http://cslt.riit.tsinghua.edu.cn/mediawiki/images/1/1f/2014_-_Modeling_Interestingness_with_Deep_Neural_Networks_-_Report.pdf pdf]<br />
: Latest approach for LSTM + RNN DSSM model : SEMANTIC MODELLING WITH LONG-SHORT-TERM MEMORY FOR INFORMATION RETRIEVAL [http://cslt.riit.tsinghua.edu.cn/mediawiki/images/2/24/2015_-_SEMANTIC_MODELLING_WITH_LONG-SHORT-TERM_MEMORY_FOR_INFORMATION_RETRIEVAL_-_Report.pdf pdf]<br />
<br />
: 2016-04-14 : Test dssm-dnn model, code dssm-cnn model.<br />
Continue investigate deep neural question answering system.<br />
: 2016-04-13 : test dssm model, investigate deep neural question answering system.<br />
: Share theano ppt [http://cslt.riit.tsinghua.edu.cn/mediawiki/index.php/%E6%96%87%E4%BB%B6:Theano-RBM.pptx theano]<br />
: Share tensorflow ppt [http://cslt.riit.tsinghua.edu.cn/mediawiki/index.php/%E6%96%87%E4%BB%B6:Tensorflow.pptx tensorflow]<br />
: 2016-04-12 : Write done dssm tensor flow version.<br />
: 2016-04-11 : Write tensorflow toolkit ppt for intern student.<br />
: 2016-04-10 : Learn tensorflow toolkit.<br />
: 2016-04-09 : Learn tensorflow toolkit.<br />
: 2016-04-08 : Finish theano version.<br />
<br />
===Deep Poem Processing With Image (Ziwei Bai)===<br />
: 2016-04-20 :combine my program with Qixin Wang's<br />
: 2016-04-10 : web spider to catch a thousand pices of images.<br />
: 2016-04-13 :1、download theano for python2.7。 2.debug cnn.py<br />
: 2016-04-15 :web spider to catch 30 thousands pices of images and store them into a matrix<br />
: 2016-04-16 :modify the code of CNN and spider<br />
: 2016-04-17 :train convouloutional neural network<br />
<br />
===RNN Music Processing for lyric (Shiyao Li)===<br />
: 2016-04-20 : learn LSTM<br />
: 2016-04-09 : web spider to catch a thousand pieces of lyrics.<br />
: 2016-04-10 : extract the keywords in the lyrics<br />
: 2016-04-13 :Read paper Memory Network.<br />
: 2016-04-15 :read the paper Memory Network and start to understand its code<br />
: 2016-04-17 :read paper end to end memory network<br />
<br />
===RNN Key word Poem Processing (Yi Xiong)===<br />
: 2016-04-20 : learn web spider<br />
: 2016-04-09 : Database for N-Gram data storing<br />
: 2016-04-10 : dictionary stored in database , dictionary based segmentation and a simple bigram segmentation<br />
: 2016-04-13 : segmentation result analysis<br />
: 2016-04-15 :improve the simple bigram segmentation<br />
: 2016-04-16 :compare the result of bigram segmentation with dictionary segmentation<br />
: 2016-04-17 :learn python (head first 50%)<br />
<br />
===RNN Piano Processing (Jiyuan Zhang)===<br />
:2016-4-12：select appropriate midis and run rnnrbm model<br />
:2016-4-13：view rnnrbm model‘s code<br />
<br />
===Recommendation System (Tong Liu)===<br />
: 2016-04-09 : 1.read a review:Machine learning:Trends,perspectives, and prospects 2.learn python ,can operate dict and set<br />
: 2016-04-12 : 1.read paper Collaborative Deep Learning for Recommender Systems and take notes.2. learn the concepts of stacked denoising autoencoder(SDAE).<br />
: 2016-04-17 :1.allocate PuTTy and Xming 2.learn python, can operate slice and iterator 3.learn release and datasets of a paper: Collaborative Deep Learning for Recommender Systems<br />
<br />
===Question & Answering (Aiting Liu)===<br />
: 2016-04-20 : read Fader's paper ()2013 <br />
: 2016-04-15 :learn dssm and sent2vec<br />
: 2016-04-16 :try to figure out how thePARALAX dataset is constructed<br />
: 2016-04-17 :download the PARALAX dataset and turn it into what we want it to be</div>Ltyhttp://index.cslt.org/mediawiki/index.php/ScheduleSchedule2016-04-22T03:06:25Z<p>Lty：</p>
<hr />
<div>=Text Processing Team Schedule=<br />
<br />
==Members==<br />
===Former Members===<br />
* Rong Liu (刘荣) : 优酷<br />
* Xiaoxi Wang (王晓曦) : 图灵机器人<br />
* Xi Ma (马习) : 清华大学研究生<br />
* DongXu Zhang (张东旭) : --<br />
<br />
===Current Members===<br />
* Tianyi Luo (骆天一)<br />
* Chao Xing (邢超)<br />
* Qixin Wang (王琪鑫)<br />
* Yiqiao Pan (潘一桥)<br />
<br />
==Work Process==<br />
===Similar questions senetence vector training with RNN/LSTM and the attention RNN/LSTM chatting model training(Tianyi Luo)===<br />
--------------------2016-04-18<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
* Finish implementing theano version of LSTM Max margin vector training.<br />
--------------------2016-04-19<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
--------------------2016-04-20<br />
* Finish submiting the camera version paper of IJCAI 2016.<br />
* Update the version of Technical Report about Chinese Song Iambics generation.<br />
--------------------2016-04-21<br />
* Finish helping Teacher Wang to prepare for text group's presentation(Tang poetry and Songci generation and Intelligent QA system) for Tsinghua University's 105 anniversary.<br />
* Submit our IJCAI paper to arxiv. (Solve a big problem about submitting the paper including Chinese chacracters)<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
<br />
===Reproduce DSSM Baseline (Chao Xing)===<br />
: 2016-04-20 : Find reproduced DSSM model's bug, fix it.<br />
: 2016-04-19 : Code mixture data model by less memory dependency done. Test it's performance.<br />
: 2016-04-18 : Code mixture data model.<br />
: 2016-04-16 : Code mixture data model, but face to memory error. Dr. Wang help me fix it.<br />
: 2016-04-15 : Share Papers. Investigation a series of DSSM papers for future work. And show our intern students how to do research.<br />
: Original DSSM model : Learning Deep Structured Semantic Models for Web Search using Clickthrough Data [http://cslt.riit.tsinghua.edu.cn/mediawiki/images/4/45/2013_-_Learning_Deep_Structured_Semantic_Models_for_Web_Search_using_Clickthrough_Data_-_Report.pdf pdf]<br />
: CNN based DSSM model : A Latent Semantic Model with Convolutional-Pooling Structure for Information Retrieval [http://cslt.riit.tsinghua.edu.cn/mediawiki/images/b/b7/2014_-_A_Latent_Semantic_Model_with_Convolutional-Pooling_Structure_for_Information_Retrieval_-_Report.pdf pdf]<br />
: Use DSSM model for a new area : Modeling Interestingness with Deep Neural Networks [http://cslt.riit.tsinghua.edu.cn/mediawiki/images/1/1f/2014_-_Modeling_Interestingness_with_Deep_Neural_Networks_-_Report.pdf pdf]<br />
: Latest approach for LSTM + RNN DSSM model : SEMANTIC MODELLING WITH LONG-SHORT-TERM MEMORY FOR INFORMATION RETRIEVAL [http://cslt.riit.tsinghua.edu.cn/mediawiki/images/2/24/2015_-_SEMANTIC_MODELLING_WITH_LONG-SHORT-TERM_MEMORY_FOR_INFORMATION_RETRIEVAL_-_Report.pdf pdf]<br />
<br />
: 2016-04-14 : Test dssm-dnn model, code dssm-cnn model.<br />
Continue investigate deep neural question answering system.<br />
: 2016-04-13 : test dssm model, investigate deep neural question answering system.<br />
: Share theano ppt [http://cslt.riit.tsinghua.edu.cn/mediawiki/index.php/%E6%96%87%E4%BB%B6:Theano-RBM.pptx theano]<br />
: Share tensorflow ppt [http://cslt.riit.tsinghua.edu.cn/mediawiki/index.php/%E6%96%87%E4%BB%B6:Tensorflow.pptx tensorflow]<br />
: 2016-04-12 : Write done dssm tensor flow version.<br />
: 2016-04-11 : Write tensorflow toolkit ppt for intern student.<br />
: 2016-04-10 : Learn tensorflow toolkit.<br />
: 2016-04-09 : Learn tensorflow toolkit.<br />
: 2016-04-08 : Finish theano version.<br />
<br />
===Deep Poem Processing With Image (Ziwei Bai)===<br />
: 2016-04-20 :combine my program with Qixin Wang's<br />
: 2016-04-10 : web spider to catch a thousand pices of images.<br />
: 2016-04-13 :1、download theano for python2.7。 2.debug cnn.py<br />
: 2016-04-15 :web spider to catch 30 thousands pices of images and store them into a matrix<br />
: 2016-04-16 :modify the code of CNN and spider<br />
: 2016-04-17 :train convouloutional neural network<br />
<br />
===RNN Music Processing for lyric (Shiyao Li)===<br />
: 2016-04-20 : learn LSTM<br />
: 2016-04-09 : web spider to catch a thousand pieces of lyrics.<br />
: 2016-04-10 : extract the keywords in the lyrics<br />
: 2016-04-13 :Read paper Memory Network.<br />
: 2016-04-15 :read the paper Memory Network and start to understand its code<br />
: 2016-04-17 :read paper end to end memory network<br />
<br />
===RNN Key word Poem Processing (Yi Xiong)===<br />
: 2016-04-20 : learn web spider<br />
: 2016-04-09 : Database for N-Gram data storing<br />
: 2016-04-10 : dictionary stored in database , dictionary based segmentation and a simple bigram segmentation<br />
: 2016-04-13 : segmentation result analysis<br />
: 2016-04-15 :improve the simple bigram segmentation<br />
: 2016-04-16 :compare the result of bigram segmentation with dictionary segmentation<br />
: 2016-04-17 :learn python (head first 50%)<br />
<br />
===RNN Piano Processing (Jiyuan Zhang)===<br />
:2016-4-12：select appropriate midis and run rnnrbm model<br />
:2016-4-13：view rnnrbm model‘s code<br />
<br />
===Recommendation System (Tong Liu)===<br />
: 2016-04-09 : 1.read a review:Machine learning:Trends,perspectives, and prospects 2.learn python ,can operate dict and set<br />
: 2016-04-12 : 1.read paper Collaborative Deep Learning for Recommender Systems and take notes.2. learn the concepts of stacked denoising autoencoder(SDAE).<br />
: 2016-04-17 :1.allocate PuTTy and Xming 2.learn python, can operate slice and iterator 3.learn release and datasets of a paper: Collaborative Deep Learning for Recommender Systems<br />
<br />
===Question & Answering (Aiting Liu)===<br />
: 2016-04-20 : read Fader's paper ()2013 <br />
: 2016-04-15 :learn dssm and sent2vec<br />
: 2016-04-16 :try to figure out how thePARALAX dataset is constructed<br />
: 2016-04-17 :download the PARALAX dataset and turn it into what we want it to be</div>Ltyhttp://index.cslt.org/mediawiki/index.php/ScheduleSchedule2016-04-22T03:05:59Z<p>Lty：</p>
<hr />
<div>=Text Processing Team Schedule=<br />
<br />
==Members==<br />
===Former Members===<br />
* Rong Liu (刘荣) : 优酷<br />
* Xiaoxi Wang (王晓曦) : 图灵机器人<br />
* Xi Ma (马习) : 清华大学研究生<br />
* DongXu Zhang (张东旭) : --<br />
<br />
===Current Members===<br />
* Tianyi Luo (骆天一)<br />
* Chao Xing (邢超)<br />
* Qixin Wang (王琪鑫)<br />
* Yiqiao Pan (潘一桥)<br />
<br />
==Work Process==<br />
===Similar questions senetence vector training with RNN/LSTM and train the attention RNN/LSTM chatting model with corpus of Bing Xiao(Tianyi Luo)===<br />
--------------------2016-04-18<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
* Finish implementing theano version of LSTM Max margin vector training.<br />
--------------------2016-04-19<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
--------------------2016-04-20<br />
* Finish submiting the camera version paper of IJCAI 2016.<br />
* Update the version of Technical Report about Chinese Song Iambics generation.<br />
--------------------2016-04-21<br />
* Finish helping Teacher Wang to prepare for text group's presentation(Tang poetry and Songci generation and Intelligent QA system) for Tsinghua University's 105 anniversary.<br />
* Submit our IJCAI paper to arxiv. (Solve a big problem about submitting the paper including Chinese chacracters)<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
<br />
===Reproduce DSSM Baseline (Chao Xing)===<br />
: 2016-04-20 : Find reproduced DSSM model's bug, fix it.<br />
: 2016-04-19 : Code mixture data model by less memory dependency done. Test it's performance.<br />
: 2016-04-18 : Code mixture data model.<br />
: 2016-04-16 : Code mixture data model, but face to memory error. Dr. Wang help me fix it.<br />
: 2016-04-15 : Share Papers. Investigation a series of DSSM papers for future work. And show our intern students how to do research.<br />
: Original DSSM model : Learning Deep Structured Semantic Models for Web Search using Clickthrough Data [http://cslt.riit.tsinghua.edu.cn/mediawiki/images/4/45/2013_-_Learning_Deep_Structured_Semantic_Models_for_Web_Search_using_Clickthrough_Data_-_Report.pdf pdf]<br />
: CNN based DSSM model : A Latent Semantic Model with Convolutional-Pooling Structure for Information Retrieval [http://cslt.riit.tsinghua.edu.cn/mediawiki/images/b/b7/2014_-_A_Latent_Semantic_Model_with_Convolutional-Pooling_Structure_for_Information_Retrieval_-_Report.pdf pdf]<br />
: Use DSSM model for a new area : Modeling Interestingness with Deep Neural Networks [http://cslt.riit.tsinghua.edu.cn/mediawiki/images/1/1f/2014_-_Modeling_Interestingness_with_Deep_Neural_Networks_-_Report.pdf pdf]<br />
: Latest approach for LSTM + RNN DSSM model : SEMANTIC MODELLING WITH LONG-SHORT-TERM MEMORY FOR INFORMATION RETRIEVAL [http://cslt.riit.tsinghua.edu.cn/mediawiki/images/2/24/2015_-_SEMANTIC_MODELLING_WITH_LONG-SHORT-TERM_MEMORY_FOR_INFORMATION_RETRIEVAL_-_Report.pdf pdf]<br />
<br />
: 2016-04-14 : Test dssm-dnn model, code dssm-cnn model.<br />
Continue investigate deep neural question answering system.<br />
: 2016-04-13 : test dssm model, investigate deep neural question answering system.<br />
: Share theano ppt [http://cslt.riit.tsinghua.edu.cn/mediawiki/index.php/%E6%96%87%E4%BB%B6:Theano-RBM.pptx theano]<br />
: Share tensorflow ppt [http://cslt.riit.tsinghua.edu.cn/mediawiki/index.php/%E6%96%87%E4%BB%B6:Tensorflow.pptx tensorflow]<br />
: 2016-04-12 : Write done dssm tensor flow version.<br />
: 2016-04-11 : Write tensorflow toolkit ppt for intern student.<br />
: 2016-04-10 : Learn tensorflow toolkit.<br />
: 2016-04-09 : Learn tensorflow toolkit.<br />
: 2016-04-08 : Finish theano version.<br />
<br />
===Deep Poem Processing With Image (Ziwei Bai)===<br />
: 2016-04-20 :combine my program with Qixin Wang's<br />
: 2016-04-10 : web spider to catch a thousand pices of images.<br />
: 2016-04-13 :1、download theano for python2.7。 2.debug cnn.py<br />
: 2016-04-15 :web spider to catch 30 thousands pices of images and store them into a matrix<br />
: 2016-04-16 :modify the code of CNN and spider<br />
: 2016-04-17 :train convouloutional neural network<br />
<br />
===RNN Music Processing for lyric (Shiyao Li)===<br />
: 2016-04-20 : learn LSTM<br />
: 2016-04-09 : web spider to catch a thousand pieces of lyrics.<br />
: 2016-04-10 : extract the keywords in the lyrics<br />
: 2016-04-13 :Read paper Memory Network.<br />
: 2016-04-15 :read the paper Memory Network and start to understand its code<br />
: 2016-04-17 :read paper end to end memory network<br />
<br />
===RNN Key word Poem Processing (Yi Xiong)===<br />
: 2016-04-20 : learn web spider<br />
: 2016-04-09 : Database for N-Gram data storing<br />
: 2016-04-10 : dictionary stored in database , dictionary based segmentation and a simple bigram segmentation<br />
: 2016-04-13 : segmentation result analysis<br />
: 2016-04-15 :improve the simple bigram segmentation<br />
: 2016-04-16 :compare the result of bigram segmentation with dictionary segmentation<br />
: 2016-04-17 :learn python (head first 50%)<br />
<br />
===RNN Piano Processing (Jiyuan Zhang)===<br />
:2016-4-12：select appropriate midis and run rnnrbm model<br />
:2016-4-13：view rnnrbm model‘s code<br />
<br />
===Recommendation System (Tong Liu)===<br />
: 2016-04-09 : 1.read a review:Machine learning:Trends,perspectives, and prospects 2.learn python ,can operate dict and set<br />
: 2016-04-12 : 1.read paper Collaborative Deep Learning for Recommender Systems and take notes.2. learn the concepts of stacked denoising autoencoder(SDAE).<br />
: 2016-04-17 :1.allocate PuTTy and Xming 2.learn python, can operate slice and iterator 3.learn release and datasets of a paper: Collaborative Deep Learning for Recommender Systems<br />
<br />
===Question & Answering (Aiting Liu)===<br />
: 2016-04-20 : read Fader's paper ()2013 <br />
: 2016-04-15 :learn dssm and sent2vec<br />
: 2016-04-16 :try to figure out how thePARALAX dataset is constructed<br />
: 2016-04-17 :download the PARALAX dataset and turn it into what we want it to be</div>Ltyhttp://index.cslt.org/mediawiki/index.php/ScheduleSchedule2016-04-22T03:04:55Z<p>Lty：</p>
<hr />
<div>=Text Processing Team Schedule=<br />
<br />
==Members==<br />
===Former Members===<br />
* Rong Liu (刘荣) : 优酷<br />
* Xiaoxi Wang (王晓曦) : 图灵机器人<br />
* Xi Ma (马习) : 清华大学研究生<br />
* DongXu Zhang (张东旭) : --<br />
<br />
===Current Members===<br />
* Tianyi Luo (骆天一)<br />
* Chao Xing (邢超)<br />
* Qixin Wang (王琪鑫)<br />
* Yiqiao Pan (潘一桥)<br />
<br />
==Work Process==<br />
===Similar questions senetence vector training with RNN/LSTM(Tianyi Luo)===<br />
--------------------2016-04-18<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
* Finish implementing theano version of LSTM Max margin vector training.<br />
--------------------2016-04-19<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
--------------------2016-04-20<br />
* Finish submiting the camera version paper of IJCAI 2016.<br />
* Update the version of Technical Report about Chinese Song Iambics generation.<br />
--------------------2016-04-21<br />
* Finish helping Teacher Wang to prepare for text group's presentation(Tang poetry and Songci generation and Intelligent QA system) for Tsinghua University's 105 anniversary.<br />
* Submit our IJCAI paper to arxiv. (Solve a big problem about submitting the paper including Chinese chacracters)<br />
* Optimize theano version of Generationg the similar questions' vectors based on RNN.<br />
<br />
===Reproduce DSSM Baseline (Chao Xing)===<br />
: 2016-04-20 : Find reproduced DSSM model's bug, fix it.<br />
: 2016-04-19 : Code mixture data model by less memory dependency done. Test it's performance.<br />
: 2016-04-18 : Code mixture data model.<br />
: 2016-04-16 : Code mixture data model, but face to memory error. Dr. Wang help me fix it.<br />
: 2016-04-15 : Share Papers. Investigation a series of DSSM papers for future work. And show our intern students how to do research.<br />
: Original DSSM model : Learning Deep Structured Semantic Models for Web Search using Clickthrough Data [http://cslt.riit.tsinghua.edu.cn/mediawiki/images/4/45/2013_-_Learning_Deep_Structured_Semantic_Models_for_Web_Search_using_Clickthrough_Data_-_Report.pdf pdf]<br />
: CNN based DSSM model : A Latent Semantic Model with Convolutional-Pooling Structure for Information Retrieval [http://cslt.riit.tsinghua.edu.cn/mediawiki/images/b/b7/2014_-_A_Latent_Semantic_Model_with_Convolutional-Pooling_Structure_for_Information_Retrieval_-_Report.pdf pdf]<br />
: Use DSSM model for a new area : Modeling Interestingness with Deep Neural Networks [http://cslt.riit.tsinghua.edu.cn/mediawiki/images/1/1f/2014_-_Modeling_Interestingness_with_Deep_Neural_Networks_-_Report.pdf pdf]<br />
: Latest approach for LSTM + RNN DSSM model : SEMANTIC MODELLING WITH LONG-SHORT-TERM MEMORY FOR INFORMATION RETRIEVAL [http://cslt.riit.tsinghua.edu.cn/mediawiki/images/2/24/2015_-_SEMANTIC_MODELLING_WITH_LONG-SHORT-TERM_MEMORY_FOR_INFORMATION_RETRIEVAL_-_Report.pdf pdf]<br />
<br />
: 2016-04-14 : Test dssm-dnn model, code dssm-cnn model.<br />
Continue investigate deep neural question answering system.<br />
: 2016-04-13 : test dssm model, investigate deep neural question answering system.<br />
: Share theano ppt [http://cslt.riit.tsinghua.edu.cn/mediawiki/index.php/%E6%96%87%E4%BB%B6:Theano-RBM.pptx theano]<br />
: Share tensorflow ppt [http://cslt.riit.tsinghua.edu.cn/mediawiki/index.php/%E6%96%87%E4%BB%B6:Tensorflow.pptx tensorflow]<br />
: 2016-04-12 : Write done dssm tensor flow version.<br />
: 2016-04-11 : Write tensorflow toolkit ppt for intern student.<br />
: 2016-04-10 : Learn tensorflow toolkit.<br />
: 2016-04-09 : Learn tensorflow toolkit.<br />
: 2016-04-08 : Finish theano version.<br />
<br />
===Deep Poem Processing With Image (Ziwei Bai)===<br />
: 2016-04-20 :combine my program with Qixin Wang's<br />
: 2016-04-10 : web spider to catch a thousand pices of images.<br />
: 2016-04-13 :1、download theano for python2.7。 2.debug cnn.py<br />
: 2016-04-15 :web spider to catch 30 thousands pices of images and store them into a matrix<br />
: 2016-04-16 :modify the code of CNN and spider<br />
: 2016-04-17 :train convouloutional neural network<br />
<br />
===RNN Music Processing for lyric (Shiyao Li)===<br />
: 2016-04-20 : learn LSTM<br />
: 2016-04-09 : web spider to catch a thousand pieces of lyrics.<br />
: 2016-04-10 : extract the keywords in the lyrics<br />
: 2016-04-13 :Read paper Memory Network.<br />
: 2016-04-15 :read the paper Memory Network and start to understand its code<br />
: 2016-04-17 :read paper end to end memory network<br />
<br />
===RNN Key word Poem Processing (Yi Xiong)===<br />
: 2016-04-20 : learn web spider<br />
: 2016-04-09 : Database for N-Gram data storing<br />
: 2016-04-10 : dictionary stored in database , dictionary based segmentation and a simple bigram segmentation<br />
: 2016-04-13 : segmentation result analysis<br />
: 2016-04-15 :improve the simple bigram segmentation<br />
: 2016-04-16 :compare the result of bigram segmentation with dictionary segmentation<br />
: 2016-04-17 :learn python (head first 50%)<br />
<br />
===RNN Piano Processing (Jiyuan Zhang)===<br />
:2016-4-12：select appropriate midis and run rnnrbm model<br />
:2016-4-13：view rnnrbm model‘s code<br />
<br />
===Recommendation System (Tong Liu)===<br />
: 2016-04-09 : 1.read a review:Machine learning:Trends,perspectives, and prospects 2.learn python ,can operate dict and set<br />
: 2016-04-12 : 1.read paper Collaborative Deep Learning for Recommender Systems and take notes.2. learn the concepts of stacked denoising autoencoder(SDAE).<br />
: 2016-04-17 :1.allocate PuTTy and Xming 2.learn python, can operate slice and iterator 3.learn release and datasets of a paper: Collaborative Deep Learning for Recommender Systems<br />
<br />
===Question & Answering (Aiting Liu)===<br />
: 2016-04-20 : read Fader's paper ()2013 <br />
: 2016-04-15 :learn dssm and sent2vec<br />
: 2016-04-16 :try to figure out how thePARALAX dataset is constructed<br />
: 2016-04-17 :download the PARALAX dataset and turn it into what we want it to be</div>Ltyhttp://index.cslt.org/mediawiki/index.php/Publication-trpPublication-trp2016-04-22T02:55:20Z<p>Lty：</p>
<hr />
<div>[[文件:Aikefu.bmp|200px]]<br />
*[[媒体文件:Cslt-trp-template.pdf|TRP-20160004: A Review of Neural QA, Tianyi Luo and Dong Wang]]<br />
<br />
[[文件:simpair.png|200px]]<br />
*[[媒体文件:Trp20160003.pdf|TRP-20160003: A study of Similar Word Model for Unfrequent Word Enhancement in Speech Recognition, Xi Ma, Dong Wang and Javier Tejedor]]<br />
<br />
[[文件:low-freq.png|200px]]<br />
*[[媒体文件:How to deal with low frequency words.pdf|TRP-20160002: Low-Frequency Words Embedding, Chao Xing, Yiqiao Pan, Dong Wang]]<br />
[[文件:maxmargin.png|200px]]<br />
*[[媒体文件:Max-margin.pdf|TRP-20160001: Max-margin metric learning for speaker recognition, Lantian Li, Chao Xing, Dong Wang]]<br />
<br />
<br />
[[文件:lowv.png|200px]]<br />
*[[媒体文件:Lowv.pdf|TRP-20150033: Learning Ordered Word Representations, Xiaoxi Wang, Chao Xing, Dong Wang, Rong Liu and Yiqiao Pan]]<br />
<br />
[[文件:Adamax.png|200px]]<br />
*[[媒体文件:Adamax Online Training for Speech Recognition.pdf|TRP-20150032: Adamax Online Training for Speech Recognition, Xiangyu Zeng, Zhiyong Zhang, Dong Wang]]<br />
<br />
[[文件:Ptrnets.png|200px]]<br />
*[[媒体文件: Ptrnets.pdf|TRP-20150031: An implementation of Pointer-Networks with Extensions, Xiaoxi Wang, Dong Wang]]<br />
<br />
[[文件:dvad.png|200px]]<br />
*[[媒体文件:dvad.pdf|TRP-20150030: DNN-based Voice Activity Detection for Speaker Recognition, Fanhu Bie, Zhiyong Zhang, Dong Wang, Thomas Fang Zheng]]<br />
<br />
[[文件:uyghur.jpg|200px]]<br />
*[[媒体文件:urghur.pdf|TRP-20150029: THUYG-20：A Free Uyghur Speech Database, Askar Rozi, Shi Yin, Zhiyong Zhang, Dong Wang, Askar Hamdulla]]<br />
<br />
[[文件:nnpre.jpg|200px]]<br />
*[[媒体文件:nnpre.pdf|TRP-20150028: Knowledge Transfer Pre-training, Zhiyuang Tang, Dong Wang]]<br />
<br />
[[文件:mmargin.png|200px]]<br />
*[[媒体文件:mmargin.pdf|TRP-20150027: Max-Margin Metric Learning for Speaker Recognition, Lantian Li, Chao Xing, Dong Wang, Thomas Fang Zheng]]<br />
<br />
[[文件:binary.jpg|200px]]<br />
*[[媒体文件:binary.pdf|TRP-20150026: Binary Speaker Embedding, Lantian Li, Chao Xing, Dong Wang, Kaimin Yu, Thomas Fang Zheng]]<br />
<br />
[[文件:rnnrl.png|200px]]<br />
*[[媒体文件:rnnrl.pdf|TRP-20150025: Relation Classification via Recurrent Neural Network, Dong Xu Zhang, Dong Wang, Rong Liu]]<br />
<br />
<br />
[[文件:dplda.png|200px]]<br />
*[[媒体文件:dplda.pdf|TRP-20150024: Learning from LDA using Deep Neural Networks, Dongxu Zhang, Tianyi Luo and Dong Wang]]<br />
<br />
<br />
[[文件:jsrl.png|200px]]<br />
*[[媒体文件:jsrl.pdf|TRP-20150023: Joint Semantic Relevance Learning with Text Data and Graph Knowledge, Dongxu Zhang, Bin Yuan, Dong Wang, Rong Liu]]<br />
<br />
<br />
[[文件:listnet.png|200px]]<br />
*[[媒体文件:listnet.pdf|TRP-20150022: Stochastic Top-k ListNet, Tianyi Luo, Dong Wang, Rong Liu, Yiqiao Pan]]<br />
<br />
<br />
[[文件:segvector.png|200px]]<br />
*[[媒体文件:segvector.pdf|TRP-20150021: Improved Deep Speaker Feature Learning for Text-Dependent Speaker Recognition, Lantian Li, Yiye Lin, Zhiyong Zhang, Dong Wang]]<br />
<br />
<br />
[[文件:Vmclass.png|200px]]<br />
*[[媒体文件:Vmclass.pdf|TRP-20150020: Document Classification with Spherical Word Vectors, Yiqiao Pan, Chao Xing, Dong Wang]]<br />
<br />
<br />
[[文件:Tlearn.png|200px]]<br />
*[[媒体文件:Tlearn.pdf|TRP-20150019: Transfer Learning for Speech and Language Processing, Dong Wang and Thomas Fang Zheng]]<br />
<br />
[[文件:Songcisample.png|200px]]<br />
*[[媒体文件:Songci.pdf|TRP-20150018: Chinese Song Iambics Generation with Neural Attention-based Model, Qixin Wang, Tianyi Luo, Dong Wang, Chao Xing]]<br />
<br />
[[文件:database.jpg|200px]]<br />
*[[媒体文件:Thuyg20-sre.pdf|TRP-20150017: AN OPEN/FREE DATABASE AND BENCHMARK FOR UYGHUR SPEAKER RECOGNITION, Askar Rozi, Dong Wang, Zhiyong Zhang, Thomas Fang Zheng]]<br />
<br />
[[文件:Thchs.png|200px]]<br />
*[[媒体文件:Thchs30.pdf|TRP-20150016: THCHS-30 : A Free Chinese Speech Corpus, Dong Wang and Xuewei Zhang]]<br />
<br />
<br />
[[文件:Su.jpg|200px]]<br />
*[[媒体文件:SUSR.pdf|TRP-20150015: Improving Short Utterance Speaker Recognition by Modeling Speech Unit Classes, Chenhao Zhang Dong Wang, Lantian Li and Thomas Fang Zheng]]<br />
<br />
[[文件:dv.png|200px]]<br />
*[[媒体文件:Dvector.pdf|TRP-20150014: Deep Speaker Vectors for Semi Text-independent Speaker Verification, Lantian Li, Dong Wang, Zhiyong Zhang and Thomas Fang Zheng]]<br />
<br />
[[文件:dark.png|200px]]<br />
*[[媒体文件:Dark.pdf|TRP-20150013: Recurrent Neural Network Training with Dark Knowledge Transfer, Dong Wang, Chao Liu, Zhiyuan Tang, Zhiyong Zhang, Mengyuan Zhao]]<br />
<br />
[[文件:PBE.png|200px]]<br />
*[[媒体文件:Probabilistic_Belief_Embedding_for_Knowledge_Population_(TRP).pdf|TRP-20150012: Probabilistic Belief Embedding for Large-scale Knowledge Population. Miao Fan, Qiang Zhou, Andrew Abel, Thomas Fang Zheng and Ralph Grishman]]<br />
<br />
[[文件:fst-fw.png|200px]]<br />
*[[媒体文件:wpair.pdf|TRP-20150011: Recognize Foreign Low-Frequency Words with Similar Pairs, Xi Ma1, Xiaoxi Wang1 and Dong Wang]]<br />
<br />
[[文件:Cdae.png|200px]]<br />
*[[媒体文件:Music.pdf|TRP-20150010: Music Removal by Denoising Autoencoder in Speech Recognition. Mengyuan Zhao, Dong Wang, Zhiyong Zhang and Xuewei Zhang]]<br />
<br />
[[文件:vmfsne.png|200px]]<br />
*[[媒体文件:Cslt-trp-template-vmfsne.pdf|TRP-20150009: VMF-SNE: Embedding for Spherical Data. Mian Wang, Dong Wang]]<br />
<br />
[[文件:ros.png|200px]]<br />
*[[媒体文件:Ros.pdf|TRP-20150008: Learning Speech Rate in Speech Recognition. Xiangyu Zeng, Shi Yin.]]<br />
<br />
[[文件:Dnnvadstru.png|200px]]<br />
*[[媒体文件:DNNVADTRP.pdf|TRP-20150007: Voice Activity Detection Based on Deep Neural Networks. Shi Yin.]] ([[媒体文件:Vad.pdf|Paper submiited to Tsinghua Xuebao]])<br />
<br />
[[文件:Uyghur-training.png|200px]]<br />
*[[媒体文件:UyghurTRP.pdf|TRP-20150006: Low-resource Uyghur Acoustic Model Training based on Cross-lingual Features. Shi Yin.]]<br />
<br />
[[文件:Beam-forming.png|200px]]<br />
*[[媒体文件:Multi-Microphones_Reverberation_Cancellation_for_Distant_Speech_Recognition.pdf|TRP-20150005: Multi-Microphones_Reverberation_Cancellation_for_Distant_Speech_Recognition. Xuewei Zhang.]]<br />
<br />
[[文件:Clipping-speaker.png|200px]]<br />
*[[媒体文件:Clip.pdf|TRP-20150004: Detection and Reconstruction of Clipped Speech in Speaker Recognition. Fanhu Bie et al.]]<br />
<br />
[[文件:Semi-dynamic-embedding.png|200px]]<br />
*[[媒体文件:Taglm.pdf|TRP-20150003: Semi-Dynamic Graph Embedding for Large Scale Language Model Adaptation. Bin Yuan et al.]]<br />
<br />
[[文件:Speaker-discriminative-score.png|200px]]<br />
*[[媒体文件:DNN-based Discriminative Scoring for Speaker.pdf|TRP-20150002: DNN-based Discriminative Scoring for Speaker Recognition Based on i-vector. Jun Wang et al. ]]<br />
<br />
[[文件:Noisy-traiing.png|200px]]<br />
*[[媒体文件:Noisy Training for Deep Neural Networks in.pdf|TRP-20150001: Noisy Training for Deep Neural Networks in Speech Recognition. Shi Yin et al.]]<br />
<br />
[[文件:English-scroing.png|200px]]<br />
*[[媒体文件:AutomaticScoringforEnglishUtterances.pdf|TRP-20140001: Automatic Scoring for English Utterances. Bo Hu.]]<br />
<br />
<br />
[[媒体文件:Template.rar|CSLT-TRP latex template]]</div>Lty