FULL STORY

ICLR's New Submission Quota Rule Sparks Controversy

To combat the surge in low-quality AI papers, ICLR announced strict submission quotas. Subsequent data analysis revealed that these limits could block a substantial percentage of annual paper submissions.

2026-07-31 ~ 2026-08-01 · 2 episodes · 9 posts

Episode 1 · ICLR Introduces Author Submission Quotas, Sparking Controversy (2026-07-31, 6 posts)

Top machine learning conference ICLR has announced a strict author submission quota system starting in 2027 to combat the surge of low-quality AI papers and alleviate reviewing burdens. The new rules explicitly cap individual submission limits and restrict the number of entries from novice teams, quickly sparking academic debates over research quality and the review mechanism.

Confirmed

  • Each author may submit a maximum of 20 papers.
  • If all co-authors of a paper have never published in a top ML conference or journal, the team can submit a maximum of 1 paper per year to ICLR (meaning each paper requires at least one co-author with top-tier publishing experience).
  • This policy applies not only to ICLR 2027 but also to ICLR 2026.

Why it matters

  • Curbing academic rat-racing: The policy is widely seen as a direct response to the current proliferation of AI papers and the focus on quantity over quality, aiming to relieve the immense pressure on the conference review system.
  • Scholar feedback and controversy: Multiple scholars have expressed differing views. Alex Dimakis and others have called for the academic community to genuinely shift towards "quality over quantity"; meanwhile, a Stanford professor pointed out that the very need to introduce a quota policy to limit submissions is itself cause for concern regarding the current academic ecosystem.

Episode 2 · Analysis Shows AI Conference Submission Caps Impact Many Papers (2026-08-01, 3 posts)

Data analysis reveals that capping AI conference submissions at 5 papers per author would block at least 22% of submissions, while a 20-paper limit would affect 2.5%. Additionally, ICLR 2026 data shows single-author papers have a much lower acceptance rate of 8.8% compared to the overall average.