CVPR 2019 Accepts Record 1300 Papers

The Conference on Computer Vision and Pattern Recognition (CVPR) announced this week they have accepted 1300 research papers for CVPR 2019, which will be held June 16 – 20 in Long Beach, California. This year’s submission and acceptance totals both set records for the world’s premier computer vision conference, which had never before accepted more than 1000 papers.

With artificial intelligence research, development and deployment booming, AI academic conferences have in recent years evolved from cozy and low-key gatherings into huge productions at major event spaces featuring celebrity keynote speakers, interactive displays and workshops, entertainment and so on. CVPR submissions spiked this year to 5,165, a 56 percent increase over 2018.

As usual, conference organizers have first publicly released only the accepted papers’ IDs, which are assigned upon submission. The authors, titles, and other details will come later. The Paper ID List is here.

It is in a computer scientist’s nature to attempt to analyse any and all types of data and that’s what Georgia Tech CS PhD student Abhishek Das (@abhshkdz) did with the histogram below, which shows the number of papers accepted within each ID number range is relatively even.

Skyrocketing submission rates reflect increasing research in machine learning, but have also created challenges, particularly in academia. AI conferences are fielding an increasing number of submissions and seeing greater participation from industry and deep-pocketed companies. This has made it more difficult for humble academic researchers to stand out from the pack. As Chief Engineer of Deep Learning at Magic Leap Tomasz Malisiewicz tweeted, “A good paper must tell a memorable story to be remembered in today’s scientific ocean of ideas.”

Another issue associated with increasing paper submission volumes is the insufficient number of reviewers. CVPR and other conferences have recently come under criticism for what some regard as unqualified reviewers or unsatisfactory review processes. At present there seems to be no ideal solution for this.

A key conference quality indicator is low paper acceptance rates. The CVPR 2019 paper acceptance rate dropped to 25.2 percent from last year’s 29 percent. For perspective, the AAAI 2019 paper acceptance rate was 16.2 percent, the first time that conference’s rate had dipped below 20 percent; while NeurIPS 2018 had a 20.8 percent acceptance rate.

CVPR 2019 has not yet announced its spotlight and poster papers.

Author: Herin Zhao | Editor: Michael Sarazen