Mitsuhiko Nakamoto

I am a second-year CS Ph.D. student in the Berkeley Artificial Intelligence Research (BAIR) Lab at UC Berkeley, advised by Professor Sergey Levine.

My research goal is to develop algorithms that empower robots with both high-level dexterity and generalization capabilities, and incorporate them into everyday life. To this end, my current interests focus on data-driven approaches to solving real-world robotic tasks, which includes offline reinforcement learning (RL), online RL fine-tuning, and imitation learning.

Prior to Berkeley, I received my bachelor's degree from the University of Tokyo in March 2022, advised by Professor Yoshimasa Tsuruoka. After graduating, I also had a great opportunity to spend a couple of months in Professor Yutaka Matsuo's Lab before moving to Berkeley. I've also worked on applications of deep learning to cardiovascular medicine at the AI group of the University of Tokyo Hospital.

Email  /  Google Scholar  /  GitHub  /  Twitter

profile photo

* denotes equal contribution

Conference Papers and Pre-prints
SuSIE: Zero-Shot Robotic Manipulation with Pretrained Image-Editing Diffusion Models
Kevin Black*, Mitsuhiko Nakamoto*, Pranav Atreya, Homer Walke, Chelsea Finn, Aviral Kumar, Sergey Levine.
arxiv link / project page / code
We propose SuSIE (Subgoal Synthesis via Image Editing), a method that leverages text-to-image diffusion models (Stable Diffusion, InstructPix2Pix) for zero-shot robot planning.
Cal-QL: Calibrated Offline RL Pre-Training for Efficient Online Fine-Tuning
Mitsuhiko Nakamoto*, Yuexiang Zhai*, Anikait Singh, Max Sobol Mark, Yi Ma, Chelsea Finn, Aviral Kumar, Sergey Levine.
To appear in NeurIPS 2023
arxiv link / project page / video
We propose calibrated Q-learning (Cal-QL), a method for acquiring an offline initialization that facilitates online fine-tuning.
Pre-Training for Robots: Offline RL Enables Learning New Tasks from a Handful of Trials
Aviral Kumar*, Anikait Singh*, Frederik Ebert*, Mitsuhiko Nakamoto, Yanlai Yang, Chelsea Finn, Sergey Levine
Robotics: Science and Systems (RSS) 2023
arxiv link / project page
Workshop Papers
Unsupervised Reinforcement Learning for Partially Observable Environments Using External Memory
Mitsuhiko Nakamoto, Yoshimasa Tsuruoka.
NeurIPS 2021 Workshop on Ecological Theory of Reinforcement Learning, 2021
paper link / project page
Propose an unsupervised RL algorithm for partially observable environments. Our algorithm enables a HalfCheetah agent to run forward and backward with limited observations and without receiving any external rewards.
Self-Supervised Contrastive Learning for Electrocardiograms to Detect Left Ventricular Systolic Dysfunction
M.Nakamoto, S.Kodera, H.Takeuchi, S.Sawano, S.Katsushika, I.Komuro.
NeurIPS 2021 Workshop on Medical Imaging Meets NeurIPS, 2021
paper link
Propose a self-supervised pretraining approach to improve the performance of deep learning models that detect left ventricular systolic dysfunction from 12-lead electrocardiography data.
A GAN Based Approach to Lip-Sync 2D Cartoon Animations without Requiring Raw Cartoon Dataset
M.Nakamoto, X.Wang, T.Yamasaki.
The 11th International Workshop on Image Media Quality and its Applications (IMQA2022), 2022
Student Presentation Award
project link
Proposed a novel approach to construct a cartoon-style speaking video dataset instead of using raw cartoon dataset for training a GAN for lip-syncing 2D cartoon animations.
Interpretable Electrocardiogram Mapping to Detect Decreased Cardiac Contraction
H.Takeuchi, S.Kodera, M.Nakamoto, S.Sawano, S.Katsushika.
NeurIPS 2021 Workshop on Bridging the Gap: From Machine Learning Research to Clinical Practice, 2021
Journal Papers
Deep Learning Algorithm to Detect Cardiac Sarcoidosis From Echocardiographic Movies
S.Katsushika, S.Kodera, M.Nakamoto, K.Ninomiya, N.Kakuda, H.Shinohara, R.Matsuoka, H.Ieki, M.Uehara, Y.Higashikuni, K.Nakanishi, T.Nakao, N.Takeda, K.Fujiu, M.Daimon, J.Ando, H.Akazawa, H.Morita, I.Komuro.
In Circulation Journal, 2021
paper link
Deep learning model to detect significant aortic regurgitation using electrocardiography
S.Sawano, S.Kodera, S.Katsushika, M.Nakamoto, K.Ninomiya, H.Shinohara, Y.Higashikuni, K.Nakanishi, T.Nakao, T.Seki, N.Takeda, K.Fujiu, M.Daimon, H.Akazawa, H.Morita, I.Komuro.
In Journal of Cardiology, 2021
paper link
Automatic detection of vessel structure by deep learning using intravascular ultrasound images of the coronary arteries
H.Shinohara, S.Kodera, K.Ninomiya, M.Nakamoto, KS.Katsushika, A.Saito, S.Minatsuki, H.Kikuchi, A.Kiyosue, Y.Higashikuni, N.Takeda, K.Fujiu, J.Ando, H.Akazawa, H.Morita, I.Komuro.
In PLOS ONE, 2021
paper link
Side Projects
Hadware Applications
Teleoperated Robot
  • This is a project done in this course
  • Designed the gripper and master interface for a teleoperated robot.
  • Technologies: CoppeliaSim, Arduino, Servo Motor, 3D CAD design, 3D Printer, potentiometers
  • [Click here to see demo videos]
Remote Switch Toggler
  • A device for turning on/off the electric switch remotely using a smartphone.
  • Technologies: ESP32, Servo Motor, BLE socket programming, electronic design (EAGLE)
  • [Click here to see demo videos]
Software Applications
Flappy Chameleon (a smartphone game app)
  • [App Store Link]
  • 2017.10-2018.2
  • Planned, designed, implemented by myself.
  • Technologies: C#, Unity
Foolip (a smartphone app for managing gourmet information)
  • App Store: (now unavailable)
  • [A review blog in Japanese]
  • 2018.6-2018.12
  • Planned, designed, implemented by a team of 3 members, mainly responsible for the front-end.
  • Technologies: React Native, Ruby on Rails
I am from Yokohama, Japan. My name in Japanese is 中本光彦.
Programming Language
  • Python (JAX, PyTorch, Tensorflow)
  • C / C++ / C#
  • Javascript (React.js, React Native)
  • SQL
  • Soccer
  • Japanese Chess (Shogi)

Template Source