Title: Toward World Models: Geometry, View Synthesis, and Visual Reasoning
Time: 2026/7/29 16:30-17:30
Location: R101, CSIE
Speaker: Prof. Ming-Hsuan Yang, University of California, Merced
Host: Prof. Yung-Yu Chuang
Abstract:
Recent advances in dynamic scene reconstruction, novel-view synthesis, and multimodal reasoning are accelerating progress toward AI world models. Methods such as DUSt3R demonstrate that scene geometry and camera motion can be recovered from casually captured images and videos, while neural rendering and generative view-synthesis techniques can produce realistic observations from previously unseen viewpoints. In parallel, multimodal models are becoming increasingly capable of reasoning about spatial relationships, scene dynamics, and physical interactions. This talk will explore how these research directions are converging toward unified world models that can reconstruct, generate, and reason about dynamic environments. I will highlight recent progress in dynamic 3D and 4D reconstruction, geometry-aware visual generation, and physically grounded reasoning, and discuss the opportunities and remaining challenges for embodied AI, robotics, and interactive visua understanding.
Bio:
Ming-Hsuan Yang is a Professor at the University of California, Merced, and a Research Scientist at Google DeepMind. His research has received numerous honors, including the Google Faculty Award in 2009, NSF CAREER Award in 2012, NVIDIA Pioneer Research Awards in 2017 and 2018, and Sony Faculty Award in 2025. He has received Best Paper
Honorable Mentions at UIST 2017 and CVPR 2018, Best Student Paper Honorable Mention at ACCV 2018, Longuet-Higgins Prize at CVPR 2023, Best Paper Award at ICML 2024, and Test-of-Time Award at WACV 2025. He was recognized as a Highly Cited Researcher from 2018 through 2025. He currently serves as Associate Editor-in-Chief of IEEE Transactions on Pattern Analysis and Machine Intelligence and as an Associate Editor of the International Journal of Computer Vision and Transactions on Machine Learning Research. He is a Fellow of IEEE, ACM, AAAI, and AAAS.