Chinese researchers dominate ICLR with 51% of accepted papers
Despite NeurIPS tensions, Chinese AI papers outpace US submissions 51% to 32% at ICLR.
Chinese technology companies and researchers turned out in force at the International Conference on Learning Representations (ICLR) in Rio de Janeiro (April 23–27), despite growing US-China tensions and a recent dispute over a NeurIPS policy that seemed to bar US-sanctioned entities. Papers with contributors from mainland China and Hong Kong accounted for over 51% of the 5,355 accepted submissions, compared with just under 32% from the United States. The conference received about 19,000 submissions with an overall acceptance rate of 28%. Major Chinese firms like Alibaba and ByteDance led the submission counts, signaling that geopolitical friction has not dampened their engagement in top-tier AI research.
This strong showing comes after several Chinese professional bodies urged domestic researchers to boycott NeurIPS over a policy that initially appeared to exclude US-sanctioned entities. The NeurIPS Foundation later clarified that the policy did not target major Chinese AI companies such as Huawei, but the episode raised questions about whether Chinese institutions would reduce participation in international AI conferences. ICLR’s results suggest the opposite: Chinese researchers are deepening their footprint, leveraging global platforms to showcase advances while navigating political headwinds.
- Chinese-affiliated papers made up 51% of ICLR’s 5,355 accepted papers, vs. 32% from the US.
- ICLR received ~19,000 submissions, with a 28% acceptance rate, held in Rio de Janeiro April 23–27.
- Participation persisted despite a NeurIPS boycott call from Chinese bodies over sanctions policy; NeurIPS later clarified major Chinese firms were not barred.
Why It Matters
Chinese AI dominance at top conferences persists despite geopolitical tensions, ensuring continued influence on global AI research directions.