Original article can be found here (source): Artificial Intelligence on Medium
The accepted papers have been announced for IJCAI — PRICAI 2020. A total of 592 papers out of 4717 submissions made it to the 29th International Joint Conference on Artificial Intelligence and the 17th Pacific Rim International Conference on Artificial Intelligence. The 12.6 percent acceptance rate based on valid submissions marked a historic low for the prestigious and discerning conference.
Reflecting the ever-increasing popularity of artificial intelligence and machine learning research, IJCAI — PRICAI 2020 saw a surge of paper submissions. Arizona State University AI researcher and Professor Subbarao Kambhampati tweeted “Interesting to note that from IJCAI in 2016 (when I happened to be the program chair) to 2020, the submissions more than doubled, going up by 2423, while accepts went up by 41.”
Rising paper submissions at global AI conferences have challenged organizers to revise submission and review policies. Some of the changes have in turn sparked criticism from the research community.
Prior to yesterday’s accepted paper announcement, IJCAI 2020 had already come under fire regarding summary rejects and rebuttals. The conference introduced its summary-reject mechanism to cope with high number of submissions. An IJCAI 2020 blog post describes how, in this phase, between 7 and 10 Senior Program Committee members are asked “to have a very quick look (five to ten minutes) at a paper and vote whether yes or no it deserves to go for standard ‘full’ review process.”
During the summary-reject phase, an area chair will analyze votes and eventually decide whether to keep or discard the paper. Some in the AI research community however are pushing for greater scrutiny of such summary rejects. New York University Associate Professor and AI researcher Julian Togelius tweeted, “I’m Senior PC for IJCAI as well this year, and took part in the summary reject voting process. I found it very hard to decide which papers were worthy of review, especially if they were not exactly in my area. The choice becomes almost random.”
Togelius also discovered that three papers he and his team submitted had failed to get through the summary-reject phase. “This is unbelievable. These papers are some of my best recent work, and I’ve published papers at IJCAI before.”
This year IJCAI also strictly restricted author rebuttals, as decisions of the program committee are final and cannot be appealed.
The official list of IJCAI — PRICAI 2020 accepted papers has not yet been published.
IJCAI-PRICAI 2020 was originally scheduled for July 11 to 17 in Yokohama, Japan. Like other AI conferences amid the COVID-19 pandemic, IJCAI-PRICAI 2020 will scratch its conventional physical gathering. Organizers have not yet announced in what form the conference will take place. Synced will update readers when this information becomes available.