Mineui Hong

Mineui Hong
[Pronunciation: "mini-hong"]

Hi! I am a postdoc researcher in Auton Lab at Carnegie Mellon University, advised by Jeff Schneider. I got Ph.D. in the Robot Learning Laboratory (RLLAB) at Seoul National University, advised by Songhwai Oh. Before starting my Ph.D. course, I received a B.S. degree in Electrical and Computer Engineering from SNU in 2019.

Research Keywords: Robot Learning, Reinforcement Learning, Planning, Representation Learning

Email / CV / Google Scholar

Publications

	Conflict-Averse Gradient Aggregation for Constrained Multi-Objective Reinforcement Learning Dohyeong Kim, Mineui Hong, Jeongho Park, and Songhwai Oh ICLR (International Conference on Learning Representations), 2025 paper We propose a novel policy gradient method that treats the maximization of multiple objectives as a constrained optimization, to avoid conflicting gradients in multi-objective RL problems.
	Diffused Task-Agnostic Milestone Planner Mineui Hong, Minjae Kang, and Songhwai Oh NeurIPS (Neural Information Processing Systems), 2023 paper \| project page \| code \| video We propose a method which predicts milestones in a latent space using a diffusion model, to guide an agent to reach a distant goal by following them.
	Dynamics-Aware Metric Embedding: Metric Learning in a Latent Space for Visual Planning Mineui Hong, Kyungjae Lee, Minjae Kang, Wonsuhk Jung, and Songhwai Oh RA-L (IEEE Robotics and Automation Letters), 2022 also presented in ICRA (International Coference on Robotics and Automation), 2022 paper \| video We propose a method to learn a latent metric space that reflects dynamical relationship between image observations, to help an agent to plan a path to reach a distant goal.
	Learning Latent Dynamics from Multi-View Observations for Image-Based Control Mineui Hong and Songhwai Oh ICCAS (International Conference on Control, Automation and Systems), 2021 paper We propose a method to learn a unified latent state space model from multi-view observations to enhance the sample-efficiency for training a visual manipulation agent.
	Generalized Tsallis Entropy Reinforcement Learning and Its Application to Soft Moblie Robots Kyungjae Lee, Sungyub Kim, Sungbin Lim, Sungjoon Choi, Mineui Hong, Jae In Kim, Yong-Lae Park, and Songhwai Oh RSS (Robotics: Science and Systems), 2020 paper \| code We propose a sample-efficient RL algorithm based on Tsallis entropy maximization and its application on training controller for a tripod soft moblie robot.
	Learning to Walk a Tripod moblie Robot Using Nonlinear Soft Vibration Actuators with Entropy Adaptive Reinforcement Learning Jae In Kim, Mineui Hong, Kyungjae Lee, DongWook Kim, Yong-Lae Park, and Songhwai Oh RA-L (IEEE Robotics and Automation Letters), 2020 also presented in ICRA (International Conference on Robotics and Automation), 2020 paper \| code \| video We propose a soft moblie robot with the new type of actuator using the nonlinear stiffness characteristic of a hyperelastic material. We also propose a novel RL algorithm that adaptively adjusts exploration-exploitation trade-off, for training a controller of the sort robot.

Research Experience

	Goal-Oriented Artificial Intelligence Agents (2022 - Present) - Funded by Ministry of Science and ICT (MSIT)
	Robot Learning: Efficient, Safe, and Socially-Acceptable Machine Learning (2019 - Present) - Funded by Ministry of Science and ICT (MSIT)
	Robot Learning from Dmonstrations with Mixed Qualities (2019 - 2021) - Funded by Ministry of Science and ICT (MSIT)

Teaching Experience

Teaching Assistant
- Graduation Project (Spring 2022)
- Introduction to Intelligent Systems (Fall 2019)

Instructor
- Python Programming Tutoring in College of Natural Sciences, Seoul National University (Fall 2018)

template adapted from this awesome website
Last updated: Sep 2023