liuxien的个人博客分享 http://blog.sciencenet.cn/u/liuxien

博文

ACL 2016 随性总结

已有 5475 次阅读 2016-9-24 16:23 |系统分类:科研笔记

A Simple Review of ACL 2016

Xien Liu
Tsinghua-iFlytek Joint Laboratory for Speech Technology, Beijing, China

September 17, 2016

1

-  Conference: 54ACL柏林, Aug 7-12, 2016.

-  Paper:

    - submissions: ()稿long paper 825, short paper 463 ;

     - acceptances:longpaper231(oral/poster=116/115)andshortpaper97(oral/poster=50/47);

     - features: TACL遴选25ACLpresentation;

     - outstanding papers: 11 (9 long, 2 short);

     -  Tutorial:8个主:-

                   - Multimodal Learning and Reasoning

                   - NLP Approaches to Computational Argumentation

                   - Computer Aided Translation

                   - Semantic Representations of Word Senses and Concepts

                   - Neural Machine Translation

                   - Game Theory and Natural Language: Origin, Evolution and Processing- Understanding Short Texts

                   - MetaNet: Repository, Identification System, and Applications

-  Session:44 sessions,3,2, Parsing 4, 5(Word Meaning 2, Word Vectors 3);


Comments:

    1. :稿稿,盲目不中,稿,;

    2. tutorial 了下 Semantic Representations of Word Senses and Con-ceptsUnderstanding Short Texts两个;

    2.1) 一个,word2vecpLSA(LDA)一些,tutorial,仅仅    纳统一为context,,(),未来一个,一个;word sense ,word embeddingword senseembedding一个,ACL中也;

       2.2) Understanding Short Texts tutorial,话语(), ,(),,等等;过这tutorial(1),; 2),embedding),;

  3. Session词语NLP;词语,word meaning, word vector,word sense讨论;


2

- : 15, 场地(poster),些不便;,ACL(15),了一讨论,ACL;

- 刃剑:ACL,NLP个主;度应,

      1. ACLtop conference恐怕一个(Google deep mind, IBMWatson,Amazon, FaceBook, Microsoft Research,,workshop讨论等等,Iflytek,),一个;

     2. focus, ACLNIPS,ICML,ICRL讨论,一亮西,讨论西,他们,;


3 NN +

他们组织MAL2015ACL2016,一些,明显NNExample-based Learning with NN ,,NLP;,一个达这所承,使NN有有,,Example-basedLearning with NN + Knowledge()NLP一个;Knowledgemiddle level,是更CMUOutstandingPaper:Harnessing DeepNeural Networks with Logic Rule一个,动加


4 (Representation Learning)

       今两个workshop一个Deep Mind组织NLP1st workshop on Representation Learning for NLP组织NLP讨论,Deep Mindworkshop中于embedding一些,e.g. ”A joint Model for word embedding and word morphology”, ”Learningword representations from multiple information sources”

       另一个Omer Levy组织NLPworkshop”1st workshop on Eval-uating vector-space representations for NLP”. Omer Levy embedding习上(e.g. 15EMNLP上一best paper讨论embedding;nipsNeural Word Embeddingas Implicit Matrix Factorization讨论embedding一些,PMI),embedding习以的相embedding一些,counting-based vs. predicting-based, explicit(embedding) vs. implicit 坏在conflict,一个Omer Levy组织么一个workshop,(embedding),,流派NLP

        从今ACL,Embedding Learning,分到word sense learning, wordmeaning, concept learning; Embedding Learningdeep,一个,今为NLP历史,svm,到前10(sparserepresentation learning),deep learning, 一个从线ICRL(International Conference Representation Learning)(恐怕很快NIPS),Google brain,Deep Mind,


5 附上几张berlin的照片吧

话说看到好几例phD带着还在推车里的baby来开会,尤其看到一个男phD在给小孩换尿不湿,很想拍下来,但是担心文化差异,没敢。 也许在计算机领域顶级会议就是这么有名魔力吧,吸引着最聪明的人到一起交流...










https://blog.sciencenet.cn/blog-832014-1004822.html

上一篇:机器学习及应用研讨会—— 个人随性总结
收藏 IP: 101.5.64.*| 热度|

0

该博文允许注册用户评论 请点击登录 评论 (0 个评论)

数据加载中...

Archiver|手机版|科学网 ( 京ICP备07017567号-12 )

GMT+8, 2024-4-26 11:56

Powered by ScienceNet.cn

Copyright © 2007- 中国科学报社

返回顶部