The International Conference on Machine Learning (ICML) is an annual academic conference dedicated to the field of machine learning. It is recognized as one of the three primary conferences of highest impact and reputation in machine learning and artificial intelligence research, alongside the Conference on Neural Information Processing Systems (NeurIPS) and the International Conference on Learning Representations (ICLR). ICML is organized by the International Machine Learning Society (IMLS). The conference typically takes place in July, with paper submissions due at the end of January, though precise dates vary by year.
History
ICML originated as the International Workshop on Machine Learning, first held at Carnegie Mellon University in Pittsburgh in July 1980. The workshop was organized by Jaime Carbonell, Ryszard S. Michalski, and Tom M. Mitchell. In 1993, the event transitioned to a full conference series under its current name, building on nearly a decade of workshops. Since the 2010s, the conference proceedings have been published open-access in the Proceedings of Machine Learning Research (PMLR), which evolved from the JMLR Workshop and Conference Proceedings and was renamed PMLR in 2015.
Notable Contributions
ICML has published numerous influential papers that have shaped modern artificial intelligence. Notable examples include the 2004 paper on K-means clustering via principal component analysis (PCA) from Lawrence Berkeley National Laboratory, the introduction of batch normalization by researchers at Google, the EfficientNet model from Google Brain (2019), the Soft Actor-Critic algorithm from University of California, Berkeley (2018), and OpenAI's CLIP model (2021). These works have had significant impact on areas such as computer vision, reinforcement learning, and representation learning.
Growth and Statistics
The conference has experienced substantial growth in submissions and acceptances. In 2015, ICML received 1,037 submissions and accepted 270 papers. By 2025, submissions had increased to 12,107, with 3,260 accepted papers. The acceptance rate has averaged between 21% (in 2020) and 30% (in 2024). In 2026, the conference set a new record with 24,371 submitted papers, more than doubling the previous year's total. Accepted papers are presented on-site as posters, with particularly outstanding contributions selected for oral presentations or designated as "spotlight papers." In 2026, 2.2% of submissions were designated as spotlights, and an additional 0.7% were selected for oral presentations. The best contribution is honored annually with the "Outstanding Paper Award" (formally the "Best Paper Award").
Scope and Sponsorship
ICML's scope spans the full breadth of machine learning, with particular emphasis on theoretical analysis, algorithmic innovation, and statistical learning. Compared to NeurIPS and ICLR, ICML traditionally features more content on statistical learning theory, reinforcement learning, robotics, and optimization theory. The conference attracts sponsors seeking access to machine learning research and talent. Technology companies such as Google, Microsoft, Amazon, Meta, and Apple regularly sponsor the event, publishing their research and recruiting researchers. Financial firms, including Citadel Securities, Jane Street Capital, and D.E. Shaw, also often have a presence at ICML.
See Also
- ICLR
- Journal of Machine Learning Research
- Machine Learning (journal)
- NeurIPS
References
- International Conference on Machine Learning official website (icml.cc)
- International Machine Learning Society website (machinelearning.org)