Hongsuk Benjamin Choi

Email: redstonepo@gmail.com

I work on robotics and reinforcement learning. My research focuses on learning robot policies through reinforcement learning and imitation learning, including generative policies for robot control.

I am a third-year PhD student in EECS at UC Berkeley and part of BAIR, advised by Prof. Angjoo Kanazawa, Prof. Jitendra Malik, and Prof. Pieter Abbeel.

Currently, I am an intern at Amazon Frontier AI and Robotics Lab. Before my PhD, I spent two years at Samsung Research in New York. See my CV (September 2026) for more information.

Google Scholar  /  GitHub  /  X

A photo of Hongsuk Choi

Updates

September 7, 2026 - ABC and Flow Policy Gradients for Robot Control have been accepted to CoRL 2026.

Publications (Selected)

ABC: Scalable Behavior Cloning with Open Data, Training, and Evaluation

Arthur Allshire*, Himanshu Gaurav Singh*, Ritvik Singh*, Adam Rashid*, Hongsuk Choi*, David McAllister*, Justin Yu, Yiyuan Chen, Huang Huang, Pieter Abbeel, Xi Chen, Rocky Duan, Phillip Isola, Jitendra Malik, Fred Shentu, Guanya Shi, Philipp Wu, Angjoo Kanazawa (* core contributors). CoRL 2026.
Homepage / Paper / Code / Dataset

Flow Policy Gradients for Robot Control

Brent Yi*, Hongsuk Choi*, Himanshu Gaurav Singh, Xiaoyu Huang, Takara E. Truong, Carmelo Sferrazza, Yi Ma, Rocky Duan, Pieter Abbeel, Guanya Shi, Karen Liu, Angjoo Kanazawa (* equal contribution). CoRL 2026.
Homepage / Paper / Code

Viser: Imperative, Web-based 3D Visualization in Python

Brent Yi, Chung Min Kim, Justin Kerr, Gina Wu, Rebecca Feng, Anthony Zhang, Jonas Kulhanek, Hongsuk Choi, Yi Ma, Matthew Tancik, Angjoo Kanazawa. RSS 2026.
Homepage / Paper

Flow Matching Policy Gradients

David McAllister*, Songwei Ge*, Brent Yi*, Chung Min Kim, Ethan Weber, Hongsuk Choi, Haiwen Feng, Angjoo Kanazawa (* equal contribution). ICLR 2026.
Homepage / Paper / Code
Visual imitation enables contextual humanoid control

Visual imitation enables contextual humanoid control

Arthur Allshire*, Hongsuk Choi*, Junyi Zhang*, David McAllister*, Anthony Zhang, Chung Min Kim, Trevor Darrell, Pieter Abbeel, Jitendra Malik, Angjoo Kanazawa (* equal contribution). CoRL 2025 (Oral, Best Student Paper Award).
Homepage / Paper / arXiv / Code

PyRoki: A Modular Toolkit for Robot Kinematic Optimization

Chung Min Kim*, Brent Yi*, Hongsuk Choi, Yi Ma, Ken Goldberg, Angjoo Kanazawa (* equal contribution). IROS 2025.
Homepage / Paper / Code
HSfM: Reconstrucing People, Places, and Cameras

Reconstrucing People, Places, and Cameras

Lea Müller*, Hongsuk Choi*, Anthony Zhang, Brent Yi, Jitendra Malik, Angjoo Kanazawa (* equal contribution). CVPR 2025 (Highlight).
Homepage / Paper / Code
FineControlNet

Fine-level Text Control for Image Generation with Spatially Aligned Text Control Injection

Hongsuk Choi*, Isaac Kasahara*, Selim Engin, Moritz Alexander Graule, Nikhil Chavan-Dafle, Volkan Isler (* equal contribution). WACV 2025.
Homepage / Paper / Code
HandNeRF

Learning to Reconstruct Hand-Object Interaction Scene from a Single RGB Image

Hongsuk Choi, Nikhil Chavan-Dafle, Jiacheng Yuan, Volkan Isler, Hyunsoo Park. ICRA 2024.
Homepage / Paper / Video / Code
Rethinking Self-Supervised Visual Representation Learning

Rethinking Self-Supervised Visual Representation Learning in Pre-training for 3D Human Pose and Shape Estimation

Hongsuk Choi*, Hyeongjin Nam*, Taeryung Lee, Gyeongsik Moon, Kyoung Mu Lee (* equal contribution). ICLR 2023.
Paper / Video
HandOccNet

HandOccNet: Occlusion-Robust 3D Hand Mesh Estimation Network

JoonKyu Park*, Yeonguk Oh*, Gyeongsik Moon*, Hongsuk Choi, Kyoung Mu Lee (* equal contribution). CVPR 2022.
Paper / Code
TCMR

Beyond Static Features for Temporally Consistent 3D Human Pose and Shape from a Video

Hongsuk Choi, Gyeongsik Moon, Ju Yong Chang, Kyoung Mu Lee. CVPR 2021.
Paper / PDF / Code / Video
Pose2Mesh

Pose2Mesh: Graph Convolutional Network for 3D Human Pose and Mesh Recovery from a 2D Human Pose

Hongsuk Choi*, Gyeongsik Moon*, Kyoung Mu Lee (* equal contribution). ECCV 2020.
Paper / PDF / Code / Video

Honors

Best Student Paper Award at CoRL 2025 for "Visual imitation enables contextual humanoid control" (2025)

SNU Distinguished Master Dissertation Award

Seoul National University Distinguished Master Dissertation Award (2022)

Experience

BAIR Logo

Ph.D. student, Electrical Engineering and Computer Sciences, UC Berkeley, Berkeley, CA, USA
2024 - present

Samsung Research America

ML Research Engineer, Samsung Research America (SRA), New York, NY, USA
2022 - 2024

Naver CLOVA AI

Visiting Researcher, Naver AI Lab, Seoul, Korea
2022

Naver Labs Europe

Research Intern, Naver Labs Europe (NLE), Grenoble, France
2021

Seoul National University

M.S., Electrical and Computer Engineering, Seoul National University (SNU), Seoul, Korea
2020 - 2022

Republic of Korea Airforce

Repulic of Korea Airforce (ROKA), Gyeryong, Korea
2015 - 2017

Seoul National University

B.S., Business | Computer Science & Engineering, Seoul National University (SNU), Seoul, Korea
2013 - 2020