Jigang Kim

Integrated M.S./Ph.D. Student at Seoul National University Department of Mechanical & Aerospace Engineering

Thirty-sixth Conference on Neural Information Processing Systems (NeurIPS) 2022: DHRL: A Graph-Based Approach for Long-Horizon and Sparse Hierarchical Reinforcement Learning

[Paper] [Bibtex] [Code]

Hierarchical Reinforcement Learning (HRL) has made notable progress in complex control tasks by leveraging temporal abstraction. However, previous HRL algorithms often suffer from serious data inefficiency as environments get large. The extended components, i.e., goal space and length of episodes, impose a burden on either one or both high-level and low-level policies since both levels share the total horizon of the episode. In this paper, we present a method of Decoupling Horizons Using a Gr...

IEEE Robotics and Automation Letters (RA-L) 2022: Unsupervised Reinforcement Learning for Transferable Manipulation Skill Discovery

[Paper] [Video] [Bibtex]

Current reinforcement learning (RL) in robotics often experiences difficulty in generalizing to new downstream tasks due to the innate task-specific training paradigm. To alleviate it, unsupervised RL, a framework that pre-trains the agent in a task-agnostic manner without access to the task-specific reward, leverages active exploration for distilling diverse experience into essential skills or reusable knowledge. For exploiting such benefits also in robotic manipulation, we propose an unsupe...

IEEE Robotics and Automation Letters (RA-L) 2022: Automating Reinforcement Learning with Example-based Resets

[Paper] [Video] [Bibtex] [Code]

Deep reinforcement learning has enabled robots to learn motor skills from environmental interactions with minimal to no prior knowledge. However, existing reinforcement learning algorithms assume an episodic setting, in which the agent resets to a fixed initial state distribution at the end of each episode, to successfully train the agents from repeated trials. Such reset mechanism, while trivial for simulated tasks, can be challenging to provide for real-world robotics tasks. Resets in robot...

International Journal of Control, Automation and Systems (IJCAS) 2022: (submitted) Distributed multi-agent target search and tracking with Gaussian process and reinforcement learning

[Paper] [Video] [Bibtex]

Deploying multiple robots for target search and tracking has many practical applications, yet the challenge of planning over unknown or partially known targets remains difficult to address. With recent advances in deep learning, intelligent control techniques such as reinforcement learning have enabled agents to learn autonomously from environment interactions with little to no prior knowledge. Such methods can address the exploration-exploitation tradeoff of planning over unknown targets in ...

International Journal of Control, Automation and Systems (IJCAS) 2021: Spatio-semantic Task Recognition: Unsupervised Learning of Task-discriminative Features for Segmentation and Imitation

[Paper] [Bibtex]

Discovering task subsequences from a continuous video stream facilitates a robot imitation of sequential tasks. In this research, we develop unsupervised learning of the task subsequences which does not require a human teacher to give the supervised label of the subsequence. Task-discriminative feature, in the form of sparsely activated cells called task capsules, is proposed for self-training to preserve spatio-semantic information of a visual input. The task capsules are sparsely and exclus...

2020 20th International Conference on Control, Automation and Systems (ICCAS) 2020: Zero-Shot Transfer Learning of a Throwing Task via Domain Randomization

[Paper] [Bibtex]

Deep reinforcement learning (DRL) on continuous robot control has received a wide range of interests over the last decade. Collecting data directly from real robots results in high sample complexities and can cause safety accidents, so simulators are widely used as efficient alternatives for real robots. Unfortunately, policies trained in the simulation cannot be directly transferred to real-world robots due to a mismatch between the simulation and the reality, which is referred to as ‘realit...

IEEE Robotics and Automation Letters (RA-L) 2020: Learning Transformable and Plannable se(3) Features for Scene Imitation of a Mobile Service Robot

[Paper] [Video] [Bibtex] [Code]

Deep neural networks facilitate visuosensory inputs for robotic systems. However, the features encoded in a network without specific constraints have little physical meaning. In this research, we add constraints on the network so that the trained features are forced to represent the actual twist coordinates of interactive objects in a scene. The trained coordinates describe 6d-pose of the objects, and SE(3) transformation is applied to change the coordinate system. This algorithm is developed...

2019 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) 2019: Fast and Safe Policy Adaptation via Alignment-based Transfer

[Paper] [Video] [Bibtex]

Applying deep reinforcement learning to physical systems, as opposed to learning in simulation, presents additional challenges in terms of sample efficiency and safety. Collecting large amounts of hardware demonstration data is time-consuming and the exploratory behavior of reinforcement learning algorithms may lead the system into dangerous states, especially during the early stages of training. To address these challenges, we apply transfer learning to reuse a previously learned policy inst...

Journal of the Korean Astronomical Society (JKAS) 2014: Misclassified type 1 AGNs in the local universe

[Paper] [Bibtex]

We search for misclassified type 1 AGNs among type 2 AGNs identified with emission line flux ratios, and investigate the properties of the sample. Using 4,113 local type 2 AGNs at 0.02<z<0.05 selected from Sloan Digital Sky Survey Data Release 7, we detected a broad component of the Ha line with a Full-Width at Half-Maximum (FWHM) ranging from 1,700 to 19,090 km/s for 142 objects, based on the spectral decomposition and visual inspection. The fraction of the misclassified type 1 AGNs am...