<2025 Fall Colloquium #2: Advances and Applications of AI Technology for Image Recognition (Prof. Junseok Lee)>

ㅇ Speaker: Prof. Joonseok Lee (Department of Data Science)

ㅇ Date & Time: Thursday, December 4, 2025, 4:00 PM

ㅇ Location: 1st Floor Lecture Room, Haedong Advanced Engineering Building (Bldg. 303)

ㅇ Lecture Overview

With the advancement of deep learning and, more recently, generative models, today's AI systems have come to understand and use language in ways strikingly similar to humans.

Recently, AI's ability to understand and represent multimedia data — not just text, but also images and video — in a human-like manner has been growing rapidly.

This lecture goes beyond simple object recognition or scene classification to introduce the speaker's recent research addressing problems that require higher-order reasoning.

Specifically, it covers referring image segmentation, which locates a target specified in natural language within a given image; video summarization, which condenses a long video into a short one capturing its key content; and image editing, which edits a given image based on natural-language instructions — discussing the technical challenges involved in each and the ideas proposed to address them.

ㅇ Speaker Bio

Building on ten years of experience researching video data analysis and recommendation systems at Google's headquarters research lab, Professor Joonseok Lee now conducts research on visual representation learning, generative models, and recommendation systems in the Department of Data Science. 

ㅇ Watch the lecture: https://www.youtube.com/watch?v=DHAsXcdNd3Y