The 36th Annual Conference on Learning Theory (COLT 2023) was held at The LaLiT Ashok, Bengaluru from 12–15 July 2023. COLT, formerly known as the conference on Computational Learning Theory, has been held annually since 1988 and for the first time in India this year. It has become the leading international conference on learning theory by maintaining a highly selective process for submissions. It is committed to high-quality articles in all the theoretical aspects of machine learning and related topics. The Proceedings of COLT are automatically accepted in the Proceedings of Machine Learning Research if the authors desire to publish a conference paper.
This year’s conference program featured 2 plenary speakers – Professors Tong Zhang from The Hong Kong University of Science and Technology (HKUST) and Asu Ozdaglar from the Massachusetts Institute of Technology (MIT), a total of 165 scientific paper presentations, and a panel discussion on finding new research connections and growing one’s research sphere effectively. The conference was well attended, with over 275 registered participants from all over the globe.
The conference was followed by a one-day workshop on Learning Theory at the J N Tata auditorium, Indian Institute of Science (IISc), Bengaluru. The workshop featured many world-renowned experts in the field of online learning and optimization, giving talks about their pioneering work in the area. The event was well-attended with approximately 50 participants in all.
The distinguished guest at the workshop was Nicolò Cesa-Bianchi, Professor of Computer Science, Università degli Studi di Milano, Italy, who has written influential articles and books on online learning algorithms. The workshop celebrated his 60th birthday in the presence of his students—current and past—and his collaborators.
The workshop was also graced by other eminent scientists including Manfred Warmuth, Yoav Freund, Sebastien Bubeck, Yishay Mansour and Claudio Gentile, who spoke on a range of topics from forecasting algorithms to multi-armed bandit optimization methods. The presentations and discussions significantly enriched understanding and opened new avenues for future research and collaboration.
COLT and the associated workshop successfully facilitated the creation of a conducive platform for Indian students, early-career scientists and researchers to interact with and learn from seasoned experts in the field. This, it is hoped, not only bolstered their knowledge and expertise but also created potential opportunities for collaborative research projects in the future.