Jun Yamada

I am a DPhil student in the Applied AI Lab (A2I Lab) at the University of Oxford advised by Professor Ingmar Posner.

I was a visiting student at the University of Southern California, advised by Professor Joseph J. Lim. I did my master at University College London, supervised by Professor John Shawe-Taylor and Dr. Zafeirios Fountas. I got my bachelor degree at Keio University, advised by Professor Akito Sakurai.

        

profile photo
News
  • [July 2024] Joining Seattle Robotics Lab at NVIDIA this summer as an intern!
  • [March 2024] One paper is accepted to L4DC!
  • [Feb 2024] One paper is accepted to TMLR!
  • [Jan 2024] One paper is accepted to ICRA 2024!
  • [Oct 2023] One paper is accepted to RA-L!
  • [Sep 2023] One paper is accepted to IEEE Humanoid!
  • [Jan 2023] One paper is accepted to ICRA 2023!
  • [Jan 2022] Our TARP paper accepted to ICLR 2022!
  • [Oct 2020] Our MoPA-RL paper on accepted to CoRL 2020!
Research

I am boradly interested in rapid skill acquisition for complex manipulation tasks, with a particular focus on leveraging generative models. So far, I have been working on (1) combining motion planning and a learning-based policy for manipulation tasks in complex environments, (2) learning-based motion planning/trajectory optimisation, and (3) world modelling for efficient skill acquisition. I have also started working on bimanual manipulation because bimanual coordination is often necessary to achieve complex manipulation tasks such as assembly.

D-Cubed: Latent Diffusion Trajectory Optimisation for Dexterous Deformable Manipulation
Jun Yamada, Shaohong Zhong, Jack Collins, Ingmar Posner,

Project page / arXiv

Efficient Skill Acquisition for Industrial Insertion Tasks in Obstructed Environments
Jun Yamada, Jack Collins, Ingmar Posner,
2024, Learning for Dynamics & Control Conference (L4DC)
Project page / arXiv

World Models via Policy-Guided Trajectory Diffusion
Marc Rigter, Jun Yamada, Ingmar Posner,
Transactions on Machine Learning Research (TMLR)
arXiv

TWIST: Teacher-Student World Model Distillation for Efficient Sim-to-Real Transfer
Jun Yamada, Marc Rigter, Jack Collins, Ingmar Posner,
2024 IEEE International Conference on Robotics and Automation (ICRA)
Project page / arXiv

RAMP: A Benchmark for Evaluating Robotic Assembly Manipulation and Planning
Jack Collins*, Mark Robson*, Jun Yamada*, Mohan Sridharan, Karol Janik, Ingmar Posner,
Robotics and Automation Letters (RAL) with presentation at the International Conference of Robotics and Automation (ICRA), 2024
Project page / arXiv

LfDT: Learning Dual-Arm Manipulation from Demonstration Translated from a Human and Robotic Arm
Masato Kobayashi*, Jun Yamada*, Masashi Hamaya, Kazutoshi Tanaka,
2023 IEEE-RAS International Conference on Humanoid Robots
Project page / arXiv

Work done during internship at OMRON SINIC X Corp.

From Primates to Robots: Emerging Oscillatory Latent-Space Dynamics for Sensorimotor Control
Alexander Luis Mitchell, Oiwi Parker Jone, Jun Yamada, Wolfgang Merkt, Ioannis Havoutis, Ingmar Posner,
2023 Conference on Cognitive Computational Neuroscience (CCN),

Leveraging Scene Embeddings for Gradient-Based Motion Planning in Latent Space
Jun Yamada*, Chia-Man Hung*, Jack Collins, Ioannis Havoutis, Ingmar Posner,
2023 IEEE International Conference on Robotics and Automation (ICRA),
The 4th UK Robot Manipulation Workshop, 2023 (Best Poster Award)
Project page / arXiv

Task-Induced Representation Learning
Jun Yamada, Karl Pertsch, Anisha Gunjal, Joseph J Lim,
International Conference on Learning Representations (ICLR), 2022
Project page / OpenReview / arXiv

We evaluate the effectiveness of representation learning approaches on visually complex environments with substantial distractors. We compare common unsupervised representation learning approaches to task-induced representations, that leverage task information from prior tasks to learn what parts of the scene are important to model and what parts can be ignored.

Motion Planner Augmented Reinforcement Learning for Robot Manipulation in Obstructed Environments
Jun Yamada*, Youngwoon Lee*, Gautam Salhotra, Karl Pertsch,
Max Pflueger, Gaurav S Sukhatme, Joseph J Lim, Peter Englert,
Conference on Robot Learning (CoRL), 2020
Project page / arXiv

Our approach augments model-free RL agents with motion planning capabilities, enabling them to solve long-horizon manipulation tasks in cluttered environments.

Evolution of a Complex Predator-Prey Ecosystem on Large-scaleMulti-Agent Deep Reinforcement Learning
Jun Yamada, John Shawe-Taylor, Zafeirios Fountas
International Joint Conference on Neural Networks (IJCNN), 2020
arXiv