Jung-Su Ha

Hi! I'm a postdoc at TU Berlin in the Learning and Intelligence Systems Lab led by Prof. Marc Toussaint.

My primary research interest includes the challenging robot manipulation planning and control problems, major difficulties of which arise from the robot's high degree of freedom, perception from raw sensory inputs, uncertainty in dynamics, etc. The goal of my research is to develop a synthesized decision-making framework that enables a robot to flexibly reason about its long-term behavior and execute the resulting plans robustly, only from the raw visual and tactile sensing. To achieve this, I enjoy integrating powerful and sophisticated techniques from the areas of mechanics, control theory, optimization as well as machine learning, and constructing unifying views across them.

Selected work

Deep Visual Constraints: Neural Implicit Models for Manipulation Planning from Visual Input

Jung-Su Ha, Danny Driess, Marc Toussaint

Preprint, 2022

[PDF] [Project Page]


Learning Geometric Reasoning and Control for Long-Horizon Tasks from Visual Input

Danny Driess*, Jung-Su Ha*, Russ Tedrake, and Marc Toussaint

IEEE Int. Conf. on Robotics and Automation (ICRA), 2021

*Equal contribution, [PDF] [Video]


Distilling a Hierarchical Policy for Planning and Control via Representation and Reinforcement Learning

Jung-Su Ha*, Young-Jin Park*, Hyeok-Joo Chae, Soon-Seo Park, and Han-Lim Choi

IEEE Int. Conf. on Robotics and Automation (ICRA), 2021

*Equal contribution, [PDF] [Video]


Describing Physics For Physical Reasoning: Force-based Sequential Manipulation Planning

Marc Toussaint, Jung-Su Ha, Danny Driess

IEEE Robotics and Automation Letters (RA-L), 2020

[PDF] [Video]

Deep Visual Reasoning: Learning to Predict Action Sequences for Task and Motion Planning from an Initial Scene Image

Danny Driess, Jung-Su Ha, Marc Toussaint

Robotics: Science and Systems (R:SS), 2020

[PDF] [Video]

A Probabilistic Framework for Constrained Manipulations and Task and Motion Planning under Uncertainty

Jung-Su Ha, Danny Driess, Marc Toussaint

IEEE Int. Conf. on Robotics and Automation (ICRA), 2020

[PDF] [Video]

Adaptive Path-Integral Autoencoders: Representation Learning and Planning for Dynamical Systems

Jung-Su Ha, Young-Jin Park, Hyeok-Joo Chae, Soon-Seo Park, Han-Lim Choi

Advances in Neural Information Processing Systems (NeurIPS), 2018

[PDF] [Video]