Jump to content

International Conference on Machine Learning

From Wikipedia, the free encyclopedia
International Conference on Machine Learning
AbbreviationICML
DisciplineMachine learning, artificial intelligence, feature learning
Publication details
History1980–present
FrequencyAnnual
yes (on openreview.net)
Websitehttps://icml.cc/

The International Conference on Machine Learning (ICML) is an international academic conference in machine learning held annually since 1980.[1] It is the oldest and, along with NeurIPS and ICLR, one of the three primary conferences of highest impact and reputation in machine learning and artificial intelligence research.[2] It is organized by the International Machine Learning Society (IMLS). Precise dates vary year to year, but paper submissions are generally due at the end of January, and the conference is generally held the following July.

History

[edit]

ICML began as the International Workshop on Machine Learning, first held at Carnegie Mellon University in Pittsburgh in July 1980 and organized by Jaime Carbonell, Ryszard S. Michalski, and Tom M. Mitchell. In 1993 it transitioned to a full conference series under the current name, building on nearly a decade of workshops.  Since the 2010s its proceedings have been published open-access in the Proceedings of Machine Learning Research (PMLR), which grew out of JMLR’s Workshop & Conference Proceedings and was renamed PMLR in 2015.[1][3]

ICML has published many important research papers that form the foundation of modern Artificial Intelligence. Some notable works include Lawrence Berkeley National Laboratory's K-means clustering via PCA (ICML 2004); Google's Batch Normalization; Google Brain's EfficientNet (ICML 2019); OpenAI's CLIP (ICML 2021) and the UC Berkeley's Soft Actor-Critic (ICML 2018).

The number of submitted (and accepted) papers has sharply increased within 10 years; from 1037 (270) in 2015 to 12107 (3260) in 2025. Meanwhile, the acceptance rate continues to averages between 21 % (2020) and 30 % (2024).[4] 2026 marked a new record with 24,371 submitted papers, more than doubling the previous year.[5] Accepted submissions are presented on-site as posters; furthermore, particularly outstanding contributions are presented orally or are designated by a committee as "spotlight papers." In 2026, 2.2% of submitted contributions were designated as "spotlight" and 0.7% additionally as "oral".[6][7]

Profile

[edit]

ICML’s scope spans the full breadth of machine learning, with particular emphasis on theoretical analysis, algorithmic innovation, and statistical learning. Compared to the other two big machine learning conferences NeurIPS and ICLR, ICML traditionally features more content on statistical learning theory, reinforcement learning and robotics, and optimization theory. Its annual proceedings are published in the open-access Proceedings of Machine Learning Research (PMLR).

ICML attracts sponsors seeking access to ML research and talent. Technology companies such as Google, Microsoft, Amazon, Meta, Apple are among sponsors, regularly publishing their research and recruiting researchers at the conference. Financial firms, such as Citadel Securities, Jane Street Capital and D.E. Shaw, also often have a presence at ICML.

Locations

[edit]
YearLocationCountrySubmissions[8]AcceptedAcceptance rate
2026Seoul[9] South Korea23,918[6]635226.6%
2025Vancouver Canada12,107326026.93%
2024Vienna Austria9653294430.50%
2023Honolulu United States6538182827.96%
2022Baltimore United States5630123321.90%
2021Vienna (virtual) Austria5513118321.46%
2020Vienna (virtual) Austria4990108421.72%
2019Los Angeles United States342477322.58%
2018Stockholm Sweden247362125.11%
2017Sydney Australia167643425.89%
2016New York City United States—322—
2015Lille France103727026.04%
2014Beijing China—310—
2013Atlanta United States—283—
2012Edinburgh United Kingdom———
2011Bellevue United States———
2010Haifa Israel———
2009Montréal Canada———
2008Helsinki Finland———
2007Corvallis United States———
2006Pittsburgh United States———
2005Bonn Germany———
2004Banff Canada———
2003Washington, D.C. United States———
2002Sydney Australia———
2001Williamstown United States———
2000Stanford United States———
1999Bled Slovenia———
1998Madison United States———
1997Nashville United States———
1996Bari Italy———
1995Tahoe City United States———
1994New Brunswick United States———
1993Amherst United States———
1992Aberdeen United Kingdom———
1991Evanston United States———
1990Austin United States———
1989Ithaca United States———
1988Ann Arbor United States———
1987Irvine United States———
1985Skytop United States———
1983Monticello United States———
1980Pittsburgh United States———

See also

[edit]

References

[edit]
  1. 1 2 "Past ICMLs Online Proceedings". University of Trier. Aktuálně.cz. 1 June 1980. Archived from the original on 2008-12-12. Retrieved 31 January 2023.
  2. ↑ "Artificial Intelligence - Google Scholar Metrics". 2020-10-07. Archived from the original on 2020-10-07. Retrieved 2020-10-07.
  3. ↑ "ICML 2008, Helsinki, Finland". University of Trier. 1 June 1980. Retrieved 31 January 2023.
  4. ↑ "ICML Statistics - Paper Copilot". 2023-04-02. Retrieved 2026-01-31.
  5. ↑ ICML [@icmlconf] (2026-01-29). "how many submissions made it over the finish line? Answer: 24371!" (Tweet). Retrieved 2026-01-31 – via X (formerly Twitter).
  6. 1 2 ICML [@icmlconf] (2026-04-30). "Decision notifications are being released on OpenReview" (Tweet). Retrieved 2026-05-06 – via X (formerly Twitter).
  7. ↑ ICML [@icmlconf] (2026-05-25). "We have sent out notifications for the 168 oral presentations at #ICML2026" (Tweet). Retrieved 2026-05-26 – via X (formerly Twitter).
  8. ↑ "ICML Statistics". Paper Copilot. 2023-04-02. Retrieved 2026-05-06.
  9. ↑ "Conferences Future Meetings". icml.cc. Retrieved 2025-07-26.
[edit]
Retrieved from "https://en.wikipedia.org/w/index.php?title=International_Conference_on_Machine_Learning&oldid=1358596665"