Laureates of NeurIPS 2020

The accepted papers in NeurIPS 2020 have been announced. This year we have 1899 accepted papers. I have compiled the metadata of all these papers, based on which I can see the laureates of this year’s conference. To determine the laureates, for both individuals and organizations, I used the following four criteria: author contribution index, first author index, organization influence index, and organization sustainability index.

If you notice any mistake or you want to know your or your organization’s rank, please leave a comment below or email me.

Author Contribution Index (ACI)

This index calculates the total contribution of an author to his all accepted papers. The contribution of an author to a single paper is defined as the reciprocal of the number of authors of that paper. The total contribution is merely a simple sum of all contribution.

Champion (2.99): Sergey Levine

Sergey Levine is an Assistant Professor in the Department of Electrical Engineering and Computer Sciences at UC Berkeley. He focuses on the intersection between control and machine learning, with the aim of developing algorithms and techniques that can endow machines with the ability to autonomously acquire the skills for executing complex tasks.

Runner-up (2.70): Mihaela van der Schaar

Mihaela van der Schaar is the John Humphrey Plummer Professor of Machine Learning, Artificial Intelligence and Medicine at the University of Cambridge, a Fellow at The Alan Turing Institute in London, and a Chancellor’s Professor at UCLA. Her focuses on machine learning, AI and operations research for healthcare and medicine.

Third place (2.67): Stefano Ermon

Stefano Ermon is an Assistant Professor in the Department of Computer Science at Stanford University, where he is affiliated with the Artificial Intelligence Laboratory and a fellow of the Woods Institute for the Environment. He focuses on techniques for scalable and accurate inference in graphical models, statistical modeling of data, large-scale combinatorial optimization, and robust decision making under uncertainty.

Top 10

No. Author ACI
1 Sergey Levine 2.99
2 Mihaela van der Schaar 2.70
3 Stefano Ermon 2.67
4 Amit Daniely 2.50
5 Lin Yang 2.43
6 Jinwoo Shin 2.37
7 Sung Ju Hwang 2.20
8 J. Zico Kolter 2.20
9 Jun Zhu 2.03
10 Jure Leskovec 1.99

Why this matters?

Researchers with high ACI are active and productive. They play an important role in defining the research trend. They also have enough funding and are an important source of vacancies for junior researchers.

First Author Index (FAI)

This index calculates the occurrence of a researcher as the first author.

Champion (5): Ilias Diakonikolas

First author of 5 papers, Ilias Diakonikolas is a faculty member in the CS department at UW Madison. He focuses on the tradeoff between statistical efficiency, computational efficiency, and robustness for fundamental problems in statistics and machine learning.

Runners-up (4): Michal Derezinski & Ruosong Wang

First author of 4 papers, Michal Derezinski is a postdoc at the Foundations of Data Analysis institute at UC Berkeley. He focuses on sampling and optimization.

First author of 4 papers as well, Ruosong Wang is a fourth year Ph.D. student at CMU. He is proudly advised by Ruslan Salakhutdinov. He currently focuses on reinforcement learning.

Runners-up with FAI=3: 10 authors

Jack Parker-Holder, Kaiqing Zhang, Qian Lou, Ziyu Wang, Aviral Kumar, Amit Daniely, Aditya Bhaskara, Constantinos Daskalakis, Shinji Ito, Yi Hao.

Why this matters?

Researchers with high FAI are rising stars. They will become the core of their respective sub-communities and lead the future research. In time, they will grow to researchers with high ACI.

Organization Influence Index (OII)

This index calculates the influence of an organization, which is defined by the total number of papers whose partial or all authors are affiliated to that organization. An organization can be either a university, a corporate, or any other entity.

Champion (198): Google (including Google Brain)

It is not surprising that Google ranks the highest again. Here, Google and Google Brain are combined. DeepMind is not counted as a part of Google in my calculation; however, some researchers denoted their affiliation as “Google DeepMind”, which forced me to include a considerable number of papers from DeepMind into Google’s account.

Runner-up (112): MIT

MIT ranks the second highest among all organization and the highest among all universities. This is thanks partially to their high quality research and partially to the collaboration known as MIT-IBM Watson AI Lab. I double-counted papers from MIT-IBM Watson AI Lab for both MIT and IBM.

Third place (106): Stanford

Stanford is the last one whose number of papers exceeds 100.

Top 10

No. Organization OII
1 Google 198
2 MIT 112
3 Stanford 106
4 Microsoft 97
5 UCB 92
6 CMU 87
7 Tsinghua 66
8 DeepMind 65
9 Oxford 65
10 Princeton 63

It is worth noting that the 7th place, Tsinghua University from China, ranks the highest among all organization outside the US (exceeding DeepMind and Oxford in UK).

Why this matters?

Organizations with high OII are the cradle of innovation and represents the future of the country they belong to.

Organization Sustainability Index (OSI)

This index calculates the organization’s sustainability defined by the number of papers whose first author is affiliated to that organization. If the author has multiple affliation, only the first one provided by the author is counted.

Champion (73): Stanford

While Google and MIT both have higher OII than Stanford, Stanford ranks the highest in terms of OSI.

Runner-up (68): Google

The includes again both Google and Google Brain.

Third place (61): MIT

Since MIT-IBM Watson AI Lab counts as MIT in this calculation, the result can only favor MIT. Still, MIT loses to Stanford and thus its place as top university in this area.

Top 10

No. Organization OSI
1 Stanford 73
2 Google 68
3 MIT 61
4 UCB 58
5 CMU 56
6 Tsinghua 47
7 Oxford 39
8 ETH Zurich 30
9 Princeton 30
10 DeepMind 28

Contrasting this table with the one associated with OII, we have the following observation:

  • DeepMind falls even lower, from 8 to 10;
  • Microsoft falls out of top 10 (ranks 11);
  • ETH Zurich makes top 10 while it is missing in the OII table (ranks 15).

Why this matters?

Organization with high OSI have good research environment and culture. They are also sources of either excellent students or excellent mentorship (for junior researchers).

As of 19 May 2020, 181 Chinese entities are subject to licensing requirement according to US Department of Commerce. Some of them have paper accepted in NeurIPS 2020:

Entity Papers
Huawei 21
SenseTime 7
Northwestern Polytechnical University 7
Megvii Technology 3
National University of Defense Technology 2
Sichuan University 2
University of Electronic Science and Technology of China 1
Yitu Technology 1
Total 44
ByteDance 6
  • I do not need to introduce Huawei since you have all known about it through its smartphones.
  • SenseTime, Megvii, and Yitu are all AI unicorns specialized in facial recognition. They are worth \$7.5 billion, \$4 billion, and \$2.4 billion respectively. They entered the Entity List, due to the use of their technology for human rights abuses against Uyghurs in Xinjiang.
  • ByteDance, the company behind TikTok, is not in the Entity List (for now). I list it here just for reference.

For NeurIPS 2020, 2.3% of accepted papers are from the US Entity List. It is hard to say at this point what consequence it entails because we miss the big picture here.

Why this matters?

Organizations in the Entity List are subject to stagnation and even downfall. These organizations are not expected to perform equally well in future conferences.

Related post

NeurIPS 2020. Comprehensive analysis of authors, organizations, and countries by Sergei Ivanov

Written on October 31, 2020