AI773: Topics in Artificial Intelligence
- Foundational Models and Physical World Models for Multimodal AI

This course is designed for graduate students who want to understand and contribute to the two frontiers that are rapidly converging in modern AI: multimodal foundation models that perceive, reason, and generate across vision, language, and audio, and physical world models that represent, simulate, and predict the 3D/4D physical world under actions. The first seven weeks of the semester are devoted to intensive lectures, given as two series: one on multimodal representation and multimodal foundation models, and one on 3D/4D representation and generation together with a review of the recent world-model literature. The rest of the semester is a paper-driven seminar in which students present and critique papers selected from a topic pool proposed by the instructors, followed by an open discussion in which the whole class brainstorms possible future research directions. Prior knowledge of machine learning and deep learning is assumed.

Syllabus: syllabus.pdf

8/25: Welcome to AI773 Fall 2026! The syllabus is now available.
8/25: Extra enrollment will be available for up to 4 students.
8/25: The paper topic pool and the presentation/panel assignment will be announced in Week 1.

When: Friday 13:00 - 15:50, Sep. 4th to Dec. 18th, 2026

Where: Bldg. 302, Room 106

Credit: 3-3-0

Enrollment cap: 24 students (3 presentation sets per session) ~ 32 students (4 sets per session)

Language: the class material is in English, while the instruction is in Korean.

Office hour (reservation only): Friday 10:00 AM, AIIS Bldg. (연구공원) 942동 208호

* Please make a reservation by emailing us in advance.
(Staff group mail: dl_ai773_202602@navercorp.com)

[Fall 2025] AI773: Topics in Artificial Intelligence - Multimodal Generative AI Theories and Applications (website)

[Fall 2024] AI773: Topics in Artificial Intelligence - Multimodal Deep Learning Theories and Applications (website)

[Fall 2023] AI773: Topics in Artificial Intelligence - Multimodal Deep Learning Theories and Applications (website)

[Spring 2022] AI773: Special Topics in Artificial Intelligence - Deep Learning and Real-world Applications (website)

* Students without a background in multimodal learning are encouraged to skim the Fall 2025 lecture slides on multimodal representation learning, multimodal foundation models, and neural graphics before the semester begins.

This website refers to source code.