search

memect / hao

好东西传送门
1.4k stars 462 forks source link

@ AngelZywei 语义网中对于特定领域的实体和实体关系抽取（对于半结构化的文本来说）的权威文章 #100

Closed haoawesome closed 9 years ago

haoawesome commented 9 years ago

NIST TAC Knowledge Base Population (KBP2014) Entity Linking Track

Entity Linking and Wikification Reading List http://nlp.cs.rpi.edu/kbp/2014/elreading.html 这组文章你看是不是符合主题？

haoawesome commented 9 years ago

相关专家 @昊奋 @孙明明_SmarterChina

angelzou commented 9 years ago

我想找两者都涵盖的综述，以及侧重于对于某个特定领域的entity relation extraction。因为我现在看到的很多是有关无结构的文本的relation extraction，而我想找的是半结构化的文本relation extraction，也可以说relation extraction from the website。比如说，一个疾病的详细描述页面（比如感冒），就会有其属性/关系的相关描述

haoawesome commented 9 years ago

@刘知远THU 微博个人认证 ：非常喜欢哈工大HIT-SCIR @付瑞吉ricky 等在ACL 2014上的 Learning Semantic Hierarchies via Word Embeddings http://t.cn/RPmtwHp 利用word embedding技术从文本中检测词汇的上下位关系，思路简单，方法高效，论文图示赏心悦目。这应该是首个充分利用word2vec的king-queen=man-woman特性的研究工作。 http://weibo.com/1464484735/Bjj9w5Ocy

@刘知远THU: NIPS 2013上的这篇Translating Embeddings for Modeling Multi-relational Data 也是基于类似的思想，针对知识图谱的多类关系，构建 head_entity + relation = tail_entity 的优化目标，也取得很不错的效果，相信会有很多后续工作追踪这类方法。

haoawesome commented 9 years ago

@黑芝麻牛白芝麻羊 ：#PKULIS国际研讨会#Indiana University丁颖老师演讲Konwledge Graph:Mining Knowledge Transfer and Usage她团队研究思路:future search:string--entity--relation--subgragh http://t.cn/RvFAQyE

http://weibo.com/2323510192/Bd47Xlhbg

haoawesome commented 9 years ago

@王威廉 ：IMLS候任主席，CRF作者Andrew McCallum今天在CIKM介绍了其最近关于Universal Schema方面的研究。其实基本概念很简单，用矩阵的行表示entity pairs，列表示各种关系（如Freebase, TAC），然后直接跑matrix factorization的算法来做关系预测。据说结果都相当好。http://t.cn/zRbenhT http://weibo.com/1657470871/AgdkQzNCa

@鲁东东胖: 本质还是entity-relation embedding @MTPanda:所以如何拿到entity pair和relation成了如分词般的基础

haoawesome commented 9 years ago

@王利锋Fandy ：硕士时搞“实体关系抽取(Relation Extraction)”，拜读过他的文章《StatSnowball: a statistical approach to extracting entity relationships》(http://t.cn/zTeVifR )，就是大家熟知的人立方 (http://t.cn/hIhVO )。

http://weibo.com/1829662477/zwOILriog

haoawesome commented 9 years ago

LBS周刊 ：呵呵, 翻译成LBS语言:点分为四类,人(person,是主体),地点(venue),实体(entity, 是客体),事件(event,activity,比如checkin),他们构成了一个四面体的立体全互联网络(relation network,graph).所以FGS把社交网络帝国的版图扩大了,也使搜索进化了//@王淮Harry: //@ColetteLiu: 转发微博

@LBS周刊 ：今年看到的第二篇好文, 继续墙裂推荐. Graph Search是在Social Graph上进行的. Social Graph,最重要的就两样数据, 点, 如people, event, pages, objects; 线, 即点之间的关系, 比如某人对某个餐馆的like, 等等. 终于明白yelp为什么大跌了. by@王淮Harry

http://weibo.com/2614938800/zf6oDtTwK

haoawesome commented 9 years ago

http://www.cs.umd.edu/srl-book/ Introduction to Statistical Relational Learning Edited by Lise Getoor and Ben Taskar Published by The MIT Press

haoawesome commented 9 years ago

http://www.cs.ubc.ca/~hkhosrav/pub/survey.pdf A Survey on Statistical Relational Learning Hassan Khosravi and Bahareh Bina

haoawesome commented 9 years ago

http://www.cs.cmu.edu/~nbach/papers/A-survey-on-Relation-Extraction.pdf A Review of Relation Extraction

haoawesome commented 9 years ago

http://ai.cs.washington.edu/pubs/279 Identifying Relations for Open Information Extraction authors Anthony Fader, Stephen Soderland and Oren Etzioni venue Conference on Empirical Methods in Natural Language Processing year 2011

haoawesome commented 9 years ago

http://fodava.gatech.edu/files/kdd2011/KDD-2011-talk-release.pdf Relation Extraction with Relation Topics

haoawesome commented 9 years ago

http://www.cs.cmu.edu/~nbach/papers/Watson-vs-SMT.pdf A Comparison between the IBM Watson DeepQA and Statistical Machine Translation Nguyen Bach

haoawesome commented 9 years ago

So, tell me about Obama again. (Library for named entity recognitition and binary relation extraction) =

https://github.com/mit-nlp/MITIE

https://twitter.com/alfredpang/status/489385846546518016

haoawesome commented 9 years ago

http://deepdive.stanford.edu/ Probabilistc entity linking, relation extraction w/ distant supervision, etc. "DeepDive: analyze your data in depth http://bit.ly/1eqLURu ”

https://twitter.com/dvdgrs/status/431146107989917699

haoawesome commented 9 years ago

http://googleresearch.blogspot.ie/2013/12/free-language-lessons-for-computers.html Language lessons for computers: Datasets for relation extraction, entity linking, multiview learning

https://twitter.com/IAugenstein/status/407972945886072832

haoawesome commented 9 years ago

http://dymphie.com/2013/09/24/entity-network-extraction-based-on-association-finding-and-relation-extraction-tpld2013/?utm_source=twitterfeed&utm_medium=twitter

Entity Network Extraction based on Association Finding and Relation Extraction #TPDL2013: http://bit.ly/1fxqoJd https://twitter.com/Dymphie/status/382443595036631040

haoawesome commented 9 years ago

http://www.quora.com/Natural-Language-Processing/What-are-the-best-algorithms-papers-on-entity-extraction-relationship-extraction-from-text

haoawesome commented 9 years ago

https://app.yinxiang.com/shard/s12/sh/59ad7778-2c75-4dea-ab83-cbcc0faef6f1/75ac6335f8a4b646dbea9d9a96333eb7