How to consult the academic data---the beginner to analyze the CV field _ depth learning with natural language processing as an example

How to consult the academic data---the beginner to analyze the CV field with natural language processing as an example

(1) Visual field Conference: CVPR,ICCV,ECCV,ICML,ICLR, etc.

(2) There is no doubt that arxiv has become an important distribution channel for the latest progress in in-depth learning, Yoshua Bengio and other famous scholars and their team's latest research results, often first published in arxiv, and then published in the relevant top-level meetings

(3) When you need to know a field, if you can find a review of the field of the latest research, Jingxiang more. The easiest way to find it is to search for "domain name + survey/review/tutorial/summary" in Google Scholar.

If the direction of the Tainu is not related to the review, generally can also find the direction of the latest papers published, read their "related work" chapter, along with the listed references, the basic understanding of the relevant research context.

The following article is reproduced in

The article is reproduced in http://mp.weixin.qq.com/s?__biz=MzIxNzE2MTM4OA==&mid=413008841&idx=1&sn= C8062acc2ec8f804990e22d5ed22ae29&scene=0#wechat_redirect

Natural language processing and artificial intelligence are a rapidly evolving field, and we often need to read the latest papers to understand the latest developments in relevant directions. The lab once had a group of students to send an e-mail to ask me how to find academic papers, which reminds me of my new graduate students at a loss si gu situation: watching the seniors talk about the field of dynamic, but do not know how to get started. After a few years of graduate students, now can finally confidently know where to learn the latest research trends. I think this may be the common confusion of beginners, rather than just tell one person to know, it is better to write down these small knowledge, so that up-and less to take detours. Of course, this summary is just a talk, only hope that someone can get a little benefit from the personal cognitive limitations, unavoidably exhaustive, also hope that we haihan correct.

International academic organizations, academic conferences and academic papers

In the computer field, there are many international professional academic organizations, through the regular organization of academic annual meeting, report academic papers, so that scholars can more easily exchange the latest results. This article takes the Natural language processing field as an example, introduces the organization of International Academic organization and academic conference, as well as the way of finding international academic papers.

Natural language Processing (NLP) overlaps to a large extent with computational linguistics (computational linguistics,cl). Similar to other computer disciplines, NLP/CL has one of its most authoritative international professional societies, called the Association for Computational Linguistics (acl,url:http://aclweb.org/), The ACL Society hosts the most authoritative international conference in the field of NLP/CL, the ACL annual meeting, and the ACL Society holds annual meetings in North America and Europe, called Naacl and Eacl respectively. In addition, the ACL society has a number of special interest groups (special interest Groups,sigs), gathered nlp/cl different fields of scholars, nature similar to a university campus interest in the community. Among them are more well-known such as Sigdat (linguistic data and corpus-based approaches to NLP), SIGNLL (Natural Language), and so on. These sigs will also convene international academic conferences, among which the more famous are the emnlp of Sigdat (conference on empirical Methods on Natural Language Processing) and SIGNLL CONLL (Conference on Natural Language). In addition, there is a International Committee on Computational Linguistics's veteran NLP/CL academic organization, which organizes every two years a name called International Conference on The International Conference of Computational Linguistics (coling) is also an important academic meeting of NLP/CL. NLP/CL's high-level academic papers are mainly distributed in several academic conferences, such as ACL, Naacl, EMNLP and coling.

The greatest happiness of scholars in the field of NLP/CL is that the ACL Web site builds a page called ACL Anthology (url:http://aclweb.org/anthology-new/), which supports the free download of most international academic conference papers in the field, It even includes academic conferences sponsored by other organizations, such as coling, IJCNLP, etc., and supports Full-text search based on Google, which is a standing hand, NLP paper I have. Because the collection of this paper is very large, and open access, many scholars also based on its research, provides a richer search support, the specific portal can refer to the ACL anthology page search box on the right side of different search buttons.

Similar to most computer disciplines, due to the rapid development of technology, the NLP/CL field attaches more importance to the publication of academic conference papers, due to the short publication cycle, and can be communicated through meetings. Of course, NLP/CL also has its own flagship academic journals, published a lot of classical academic papers, that is computational linguistics (Url:http://www.mitpressjournals.org/loi/coli). The journal has only a few articles per period, the average quality is higher than the conference papers, time permitting words worth tracking. In addition, the ACL society has just started the transactions of ACL (tacl,url:http://www.transacl.org/) in order to enhance academic influence, which is worth paying attention to. It is worth mentioning that these two periodicals are also open access. There are also periodicals related to NLP/CL, such as ACM transactions on Speech and Language PROCESSING,ACM transactions on Asian Language information Essing,journal of quantitative linguistics et cetera.

According to the Google Scholar Metrics 2015 evaluation of NLP/CL academic journals and conferences, ACL, EMNLP, Naacl, coling, Lrec, computational linguistics are located in the top 5 places, It basically reflects the degree of concern of scholars in this field.

It is worth mentioning that the United States Hal DAUMÉIII maintained a Natural language Processing blog (http://nlpers.blogspot.com/), often commenting on the latest academic developments, deserves attention. I often read his comments on the conference of the ACL, NAACL and other academic conferences, which is very enlightening. In addition, the ACL learned to maintain a wiki page (http://aclweb.org/aclwiki/), including a large number of NLP/CL related information, such as well-known research institutions, the previous meeting rate, and so on, are necessary for home products, it is worth digging deep.

NLP/CL is the main research field of large text data intelligence. As an interdisciplinary subject, large data intelligence is closely related to the following aspects: (1) Information retrieval and data mining. The relevant academic conferences are mainly sponsored by the American Computer Society (ACM), including Sigir, WWW, WSDM, KDD and so on. Related periodicals include ACM Tois, ACM TKDD, IEEE TKDE, etc. (2) The field of artificial intelligence. Related academic conferences mainly include IJCAI and AAAI, and related academic periodicals mainly include artificial intelligence and Journal of AI Research. (3) Machine learning field, related academic conferences mainly include Icml,nips,aistats,uai, and related academic periodicals mainly include Journal of Machine Learning Research (JMLR) and Machine Learning (ML) and so on.

The Chinese Computer Society (CCF) has formulated the catalogue of international academic conferences and periodicals recommended by the China Computer Society (http://www.ccf.org.cn/sites/ccf/ paiming.jsp), the basic fair list of each field of high-level journals and conferences, has become a lot of computer systems in China's important indicators of evaluation. In the latest edition of the catalogue, which was launched at the end of 2015, the famous meetings such as ACLs and Nips were deserved. Through this list, you can quickly learn about the major journals and academic conferences in each field.

It is noteworthy that although the publication cycle of academic conference papers in the field of computer science has been very short, it is still unable to meet the rapid development in the direction of the recent in-depth study. As a result, more and more scholars choose to bypass the review process of academic conferences or periodicals and publish papers online directly via arxiv (http://arxiv.org/) and other pre-printed platforms. These latest academic results are released faster because of the lack of peer-reviewed processes. However, due to the lack of peer review of the views and filtering, resulting in the printed platform on the issue of the quality of the paper, there is a strong need to identify the ability to find a truly valuable work. Now, although natural language processing, such as the direction of the arxiv through the first release results, there are divergent views, but there is no doubt that arxiv has become an important channel for in-depth study of the latest progress, Yoshua Bengio and other famous scholars and their team's latest research results, It is often published on the arxiv and then posted at the relevant top-level meeting. Therefore, arxiv is an important information channel to understand the latest progress of large data intelligence.

Domestic academic organizations, academic conferences and academic papers

Similar to the international, there is also a NLP/CL-related academic organization in China, Chinese Information Society (url:http://www.cipsc.org.cn/). Through the Institute of the Board of Directors (http://www.cipsc.org.cn/lingdao.php) can basically understand the domestic engaged in the nlp/cl of the main units and scholars. The Chinese Information Society organizes a number of academic conferences every year, such as the National Symposium on Computational Linguistics (CCL), the Symposium on Young Scholars of natural language processing in China (YSSNLP), the National Conference on Information Retrieval (CCIR), and the National Machine Translation Seminar (CWMT), which is the domestic nlp/ An important platform for the academic exchange of CL scholars. In particular, China's natural language processing Young Scholars Seminar is dedicated to the domestic NLP/CL young scholars seminars, using the invitation system to participate, everyone volunteered to report at the Symposium on the forefront of cutting-edge academic developments, is the domestic nlp/cl young scholars to conduct academic exchanges, the establishment of an excellent platform for academic cooperation.

The 2010 coling and the 2015 ACL were held in Beijing, and the Chinese Information Society was responsible for organizing the work, which partly reflected the importance of the Institute in the field of domestic nlp/cl. In addition, the Conference on Natural Language Processing and Chinese computing (NLP&CC), organized by the Chinese Information Technology Special Committee of the Computer Society, is the most important NLP/CL academic conference in China. The Chinese information Society editor-in-chief has a long history of "Chinese Information journal", which is an important academic periodical in this field, and has published many heavy papers. In addition, the domestic famous "Computer journal", "Software Journal" and other periodicals are often published NLP/CL papers, worthy of concern.

In the past few years, the AI and NLP suites in the water wood Community BBS have been an important platform for online communication discussions in domestic NLP/CL field. In recent years with the development of social media, more and more scholars to the Sina Weibo, there is a strong communication atmosphere. How to find these scholars. An easy way is in Sina Weibo search "Find someone" function to retrieve the "natural language Processing", "Computational Linguistics", "Information Retrieval", "machine learning" and so on, immediately with the past only in the paper to see the names of teachers and students close to the exchange. There is also a way, Tsinghua University Liang Development of the "micro-Blog Tracing" System (http://xunren.thuir.org/) can retrieve the influential people in each field, so can also be used to find important scholars in the Nlp/cl field.

It is worth mentioning that many teachers and students who teach abroad are also active on Sina Weibo, such as William Wang (http://weibo.com/u/1657470871), Li Yu (http://weibo.com/mli65), and so on, often release important industry news, Worthy of concern. Weibo also has a number of special to organize and push the information related to artificial intelligence accounts, such as the Good Thing Portal (http://memect.com/), every day to organize the release of daily Domain dynamics. Academic research needs both hard skills and communication with others. The so-called speaker has no intention, the listener is intentional, perhaps other people's words can wake up your puzzling question. There is no doubt that social media such as Weibo provide a good platform for communication, but should also be careful not to indulge.

How to quickly understand the progress of a field

Finally, a quick understanding of the progress of research in a field of experience. You will find that the search engine is an important tool to look up the literature, especially Google Scholar Google, because of its huge index, will be a tool for us to hack.

When you need to know a field, it's much more Jingxiang to find a review of the latest research in the field. The easiest way to find it is to search for "domain name + survey/review/tutorial/summary" in Google Scholar. There are also a number of publishers specializing in publishing articles in various fields, such as the Foundations and trends series published by now publisher, Morgan & Claypool Publisher's synthesis lectures on Human Language Technologies series and so on. They have published a number of popular trends, such as document summaries, affective analysis and opinion mining, learning sequencing, language models, and so on.

If the direction of the Tainu is not related to the review, generally can also find the direction of the latest papers published, read their "related work" chapter, along with the listed references, the basic understanding of the relevant research context. Of course, there are many other options, such as going to the videolectures.net to see the tutorial reports made by famous academics at major academic conferences or summer schools, to directly consult researchers in this field, and so on.