Loading...

학술행사

Korean AI Association

학술행사

  >   학술행사   >   국내학술대회

국내학술대회

연사 및 초록 (26일)
 
 
최종현 교수(서울대학교)
 
Title:  Adapting Actions to Changing Environments
 
Abs: 
VLAs often fail when discrepancies arise between a pre-defined plan and the actual state of the world. To address this, I will introduce a method allowing agents to detect mismatches in object status and revise their actions before making costly mistakes. Building on this adaptive capability, I will introduce a dual-process framework that integrates continuous video monitoring with high-level strategic planning. These methods demonstrate that integrating real-time feedback and hierarchical reasoning plays a role for achieving robust, human-like adaptability in complex real-world scenarios.
 
Bio:
Jonghyun Choi received the B.S. and M.S. degrees in electrical engineering and computer science from Seoul National University, Seoul, South Korea in 2003 and 2008 respectively. He received a Ph.D. degree from University of Maryland, College Park in 2015, under the supervision of Prof. Larry S. Davis. He is currently an associate professor at Seoul National University, Seoul, South Korea. He was an associate professor at Yonsei University (2022-2024), an assistant professor at GIST (2018-2022), a research scientist at Allen Institute for Artificial Intelligence (AI2), Seattle, WA (2016-2018), a senior researcher at Comcast Applied AI Research, Washington, DC (2015-2016). He serves as an area chair at NeurIPS, ICLR, CVPR, ICCV, ECCV, WACV, AAAI (senior PC), ICRA (associate editor), and an associate editor in IEEE Transactions on PAMI and IJCV. His research interest includes visual understanding for edge devices and household robots and visual recognition of images and videos under computational and data efficiency constraints.
 
 
유현우 교수(성균관대학교)
 
Title:  Robots as Systems, Not Just AI
 
Abs: 
As robotics research increasingly emphasizes foundation models and data-driven control, platform design and manufacturing are often treated as secondary concerns. This lecture challenges that assumption.
Through case studies from electric vehicles and humanoid robotics, we show that platform simplification, manufacturing internalization, and system-level control are decisive factors for real-world deployment.
We argue that AI-centric, hardware-agnostic approaches face structural limitations in controllability and reliability, and that manufacturing-driven, system-integrated strategies provide a more sustainable path toward scalable robotic systems.
 
Bio:
Work Experience
(Mar `24 ~ Current)  Assistant Professor, Dept. of Intelligent Robotics, SKKU, South Korea
(Feb `22 ~ Feb `24)   Assistant Professor, Electrical Eng. & Grad. School of AI (affiliated), UNIST, South Korea
(Mar `20 ~ Dec `21)  Postdoctoral Researcher, Robotics Institute, Carnegie Mellon University, US
Education
Ph.D., Electrical and Computer Engineering, Seoul National University, South Korea
(Thesis: A Variational Observation Model of 3D Multi-Object in 2D Single Scene for Semantic SLAM)
B.S., Electrical and Computer Engineering, Seoul National University, South Korea
 
 
윤상웅 교수(UNIST)
 
Title:  Generative Modeling and Reinforcement Learning: Energy, Reward, and Value
 
Abs: 
In this lecture, we will investigate the deep connection between generative modeling and reinforcement learning (RL), the two pillars of machine learning. From a probabilistic viewpoint that unifies these approaches, we will discuss generative models, including energy-based models, diffusion models, and autoregressive models, from an RL perspective and the post-training of foundation models from the perspective of generative modeling.
 
Bio:
Sangwoong Yoon is an assistant professor at Graduate School of AI in Ulsan National Institute of Science and Technology (UNIST). His research focuses on the intersection between generative modeling and reinforcement learning, aiming at building generative models that can actively yet safely interact with the world.
 
 
 
주원영 교수(이화여자대학교)
 
Title:  Probabilistic Graphcial Model: Variational Inference and Structured Representation Learning
 
Abs: 
This lecture focuses on introducing probabilistic graphical models (PGM), a key framework widely used in machine learning. We will first explore variational inference, a technique for learning in PGMs, and progress to Variational Autoencoders (VAE) as a deep generative model. Additionally, we will discuss structured representation learning, building on these foundational concepts. Finally, practical applications and challenges of these models in real-world machine learning tasks will be covered.
 
Bio:
Weonyoung Joo is an assistant professor in the Department of Statistics, EWHA Womans University, Republic of Korea. He earned his Ph.D. from the Department of Industrial and Systems Engineering, Korea Advanced Institute of Science and Technology (KAIST). Prior to that, he received his M.S. and B.S. in the Department of Mathematical Sciences, KAIST. Before joining the EWHA, he served as a machine learning engineer at Samsung Electronics.