ICML 2026 Kicks Off with Record-Breaking Attendance and Submissions

The International Conference on Machine Learning (ICML) 2026 has officially opened. According to the official account, in-person attendance has surpassed 20,000, setting a new record. The rapid expansion of the conference scale is reflected not only in attendance but also in paper submissions, directly illustrating the continuous heating up of global AI and machine learning research.

Core Statistics and Acceptance

The conference received 24,661 paper submissions from 76,159 authors. After reviews by 17,886 reviewers, 1,728 area chairs, 251 senior area chairs, and 56 ethics reviewers, a total of 6,552 papers were accepted, resulting in an acceptance rate of 26.6%. Among the accepted papers, Large Language Model (LLM) related research accounted for 17.6%, maintaining its position as the most stable top-tier topic. Meanwhile, the overall proportion of AI application papers increased compared to last year, reflecting a growing academic focus on practical implementations.

Reviewing Pressure and Countermeasures

The doubling of submission volume compared to the previous edition placed immense pressure on the reviewing process. Researcher @hhsun1 pointed out that this growth rate has sparked concerns within the academic community regarding the long-term sustainability of top-tier AI conferences. To address the challenges brought by this scale growth, the ICML 2026 program committee opted against setting submission caps or charging additional fees. Instead, they introduced a new policy requiring mandatory disclosure of concurrent submissions and modified the reciprocal reviewing system, striving to maintain review quality despite the surge in submissions.

2026-07-07 ~ 2026-07-07 · 8 related posts

Primary sources