[Free Live] The process of developing an AI competition-winning algorithm into an oral presentation paper for an international conference

This is intended for developers, data scientists, AI/ML researchers, and graduate students who are participating in AI competitions or wish to develop ideas gained from competitions into actual research and papers. Starting with the time-series prediction algorithm used to win 3rd place in the K-League-University of Seoul Open AI Competition (in which 947 teams participated), we will share the actual research process—from problem analysis to algorithm design, performance improvement, research idea derivation, and paper writing. In particular, we will introduce a case-centered approach that goes beyond simply achieving good results in a competition. We will discuss how the limitations of the winning algorithm were identified and evolved into a new algorithm called FS-DCM, and the process of developing this into an Oral paper for the Methods Track at the International Conference on Automated Machine Learning (AutoML Conference 2026). Through the entire flow of "Competition → Winning Algorithm → Research Idea → Algorithm Advancement → Paper → International Conference Oral," we will explore how to connect AI competition experiences to actual research achievements.