Google, Stanford, and MIT Dominate Record-Breaking NeurIPS 2020 Accepted Papers
NeurIPS 2020 accepts 1,903 research papers at a 20.1 percent rate, with Google, Stanford, and MIT leading the institutional rankings. The United States tops the national board, while the virtual conference takes place in December.
NeurIPS 2020 releases its list of accepted papers after a record-breaking submission cycle that sees a 38 percent increase over the previous year. Out of 9,467 full submissions, 1,903 papers earn acceptance, resulting in a 20.1 percent acceptance rate that is slightly lower than last year's 21 percent. This massive influx of research highlights the growing influence and scale of the neural information processing community.
Google, Stanford, and MIT emerge as the top three institutional affiliations on the accepted papers list, contributing 169, 104, and 98 papers respectively. The United States dominates the global landscape with 1,178 accepted papers, followed by China with 259 and the United Kingdom with 195. These numbers reflect the ongoing concentration of advanced AI research within a handful of elite American universities and tech giants.
Despite the release of the list, some researchers express frustration in the Reddit machine learning community regarding the lack of proper paper links, abstracts, videos, and tags. Critics argue that this missing metadata makes it difficult to navigate the extensive list or locate specific research topics. NeurIPS 2020 proceeds as a completely virtual event, scheduled to run from December 5 through December 12.