A paper titled “Memory as Dynamics: Learning Reliability-Guided Predictive Models for Online Video Perception,” co-authored by Kim Min-woo, a senior in the Department of Software Engineering at Kookmin University (President Jeong Seung Ryul), has been accepted for presentation at ICML 2026 (Forty-Third International Conference on Machine Learning), one of the top international conferences in the field of artificial intelligence.

ICML (International Conference on Machine Learning) is recognized as the world’s most prestigious international conference in the fields of machine learning and artificial intelligence, and is considered one of the top three conferences in the AI field alongside NeurIPS and ICLR. Since its inception in 1980, it has established itself as a key platform where researchers from around the world present the latest machine learning research findings every year.

This paper presents a new framework in the field of online video perception that interprets memory not as a static repository but as a dynamic system. By introducing a reliability-guided predictive model, the research utilized temporal information within video sequences more accurately and efficiently.

In particular, by dynamically estimating the reliability of each frame and incorporating this into memory updates and predictions, the study demonstrated that stable recognition performance can be maintained even in the presence of noise or occlusion. Furthermore, it proved superior performance compared to existing state-of-the-art methods across various online video benchmarks.

This research is significant in that it presents the potential to simultaneously enhance both the accuracy and stability of video recognition. The proposed framework is expected to be applied in various fields where real-time recognition is critical, such as autonomous driving, robotics, and intelligent video understanding.
Student Kim Min-woo stated, “I sought to reinterpret the relationship between memory and prediction from a new perspective,” adding, “It is very meaningful to have the research I conducted as an undergraduate recognized at a world-class academic conference.” He added, “I hope to continue conducting practical AI research that contributes to solving real-world problems.”

This research was conducted with support from the National Research Foundation of Korea’s Mid-Career Researcher Program and the Institute for Information & Communications Technology Planning & Evaluation (IITP)’s SW-Centered University Project.

△ Kim Min-woo, a senior in the Department of Software Engineering at Kookmin University