Network Security Internet Technology Development Database Servers Mobile Phone Android Software Apple Software Computer Software News IT Information

In addition to Weibo, there is also WeChat

Please pay attention

WeChat public account

Shulou

Academic | 20-year data analysis from NIPS to NeurIP: MIT contributed the most, while Wu Enda and Hinton were the most TOP authors

2025-02-27 Update From: SLTechnology News&Howtos shulou NAV: SLTechnology News&Howtos > Internet Technology >

Share

Shulou(Shulou.com)06/03 Report--

Produced by big data Digest

Compiled by Jiang Baoshang

NeurIPS (formerly known as NIPS) is the biggest event of the year for artificial intelligence. Every year on Christmas Eve, artificial intelligence enthusiasts and scientists around the world gather to release the latest research and discuss it enthusiastically. This is not only a big party, but also an important direction of technological development. The technology of the conference will often evolve into real research and even application results in the next few years.

On the occasion of this year's NeurIPS2018, Microsoft released a study that made a comprehensive analysis of the number of papers, references, participating research institutions and authors of the conference each year since 1996 to see what interesting results there are.

Changes in the number of conference papers received

From the picture above, we can see the change in the number of conference papers from 1996 to 2017. The number of output each year is precisely marked, from 163 in 1996 to 954 in 2017, with a higher growth rate of papers in recent years. In addition, conference papers and Poster are included in the statistics.

The black bar in the above figure indicates the average number of references cited in the selected papers, and the data show that there is a positive correlation between the number of papers published and the number of references cited. The green bar shows the average number of times that conference papers are cited. Note that the reference is the original count and is not standardized by release time.

Note: average citation (Average Citations): the average number of citations of a NIPS paper received in a particular conference year.

Average references (Average References): the average number of references in a particular conference year

NeurIPS revised the submission guidelines in 2009. The increase in average references in 2009 may be due to changes in rules.

Time span of references

What is the time span of citation of NIPS papers? The above picture can give you a clear display. For example. The papers of the 2016 NIPS conference cited 2033 papers published in 2015, 1395 papers published in 2014, more data information, and so on.

Note: if some years refer to articles from future years, there are two situations that can be explained. 1. The article was later published in journals. 2. Books are quoted.

Analysis of the sources of references

What articles are cited in NIPS conference papers? What is the domain distribution of references?

The above table shows the domain distribution of references. The bar chart on the left shows the 10 areas with the largest number of references. NIPS, ICML, the Journal of Machine Learning Research (Journal of Machine Learning Research) and CVPR ranked in the top four. The pie chart on the right shows that the top 10 references account for almost 40% of the total citations from NIPS papers.

Top research institutions

How many papers received by NIPS are provided by top institutions? The bubble chart above visualizes the problem. The color depth of the bubbles is proportional to the total number of papers provided by the organization, which, of course, refer to those received by NIPS.

More detailed data

Author analysis

Who is the best author of the NIPS paper?

The next three charts rank the authors according to different criteria.

The bubble chart shows the NIPS authors ranked by number of citations, where the bubble color depth is related to the number of releases.

The bubble chart above visualizes the author's ranking. The ranking index used is calculated by Microsoft academic institutions using a formula. The X axis represents the author's ranking. The higher the level of authors, the closer they are to the right. The Y-axis controls rankings by the number of papers published. This diagram enables us to identify influential authors who have not published a large number of articles. The higher the authors are at the top of the chart, the higher their overall ranking. The author in the upper right corner of the chart not only has a large number of papers, but also has great influence.

Whose articles are most frequently cited by NIPS?

The above figure ranks the authors of references, and the more references are provided, the higher the ranking. Scholars who have not published articles in NIPS may also appear in the picture above.

The NIPS citation distribution of the top 20 scholars is updated once a year.

Related reports:

Https://www.microsoft.com/en-us/research/project/academic/articles/neurips-conference-analytics/

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.

Views: 0

*The comments in the above article only represent the author's personal views and do not represent the views and positions of this website. If you have more insights, please feel free to contribute and share.

Share To

Internet Technology

Wechat

© 2024 shulou.com SLNews company. All rights reserved.

12
Report