Andrew Sang-Jin Choi
Hi! Thanks for visiting my website ^_^
Big Update:
I recently defended my thesis titled "Simulation of Deformable Objects for Sim2Real Applications in Robotics"! 🥳🥳🥳
I am now starting a new role as a research scientist on the General AI team at Horizon Robotics.
I obtained my Ph.D. in computer science at UCLA, where I
was a member of the Structures-Computer Interaction
(SCI) Lab.
I was co-advised by Professors M. Khalid Jawed, Jungseock Joo, and Demetri Terzopoulos.
Before that, I obtained a B.S. in mechanical engineering at UC Davis and an M.S. in computer science at
UCLA.
During my Ph.D., my research was quite multidisciplinary, involving fields such as robotics,
simulation, perception, and deep learning. A key focus during my studies was the
sim2real gap. I was especially interested in leveraging physics-based simulations to teach
useful and transferable skills to robots through data-driven methods.
Moving forward, I would like to continue research in sim2real and general robotic AI.
I am particularly interested in differentiable simulations, imitation learning,
reinforcement learning, foundational models, and rapid 3D reconstruction via NeRF.
Email  / 
CV  / 
LinkedIn  / 
Google Scholar
 / 
GitHub
|
|
|
Learning Neural Force Manifolds for Sim2Real Robotic Symmetrical Paper Folding
Andrew Choi*,
Dezhong Tong*,
Demetri Terzopoulos,
Jungseock Joo,
M. Khalid Jawed
IEEE Transactions on Automation Science and Engineering (T-ASE), 2024
(* equal contribution)
Source Code
/
Video
/
Oral Presentation
/
News
Coverage
End-to-end pipeline for full sim2real robotic paper folding.
Neural force manifolds are learned from physical simulation and are used to generate optimal
folding trajectories.
Scaling analysis is used to obtain a generalizable control policy with respect to material
and
geometric parameters.
|
|
DisMech: A Discrete Differential Geometry-based
Physical Simulator for Soft Robots and Structures
Andrew Choi,
Ran Jing,
Andrew Sabelhaus,
M. Khalid Jawed
IEEE Robotics and Automation Letters (RA-L), 2024
Source Code /
Video /
Oral Presentation
Full scale generalizable physical simulator capable of simulating soft physics, frictional contact, and continuum control.
A fully implicit formulation allows for increased computational efficiency while maintaining physical accuracy when compared to previous SOTA.
|
|
Sim2Real Neural Controllers for Physics-Based Robotic Deployment of
Deformable Linear Objects
Dezhong Tong,
Andrew Choi,
Longhui Qin,
Weicheng Huang,
Jungseock Joo,
M. Khalid Jawed
The International Journal of Robotics Research (IJRR), 2023
Source Code
/
Video
End-to-end pipeline for full sim2real robotic DLO deployment.
Physical simulations are used to train a neural controller capable of deploying rods along a
prescribed pattern.
Scaling analysis is used to obtain a generalizable control policy with respect to material
and geometric parameters.
|
|
mBEST: Realtime Deformable Linear Object Detection Through Minimal
Bending
Energy Skeleton Pixel Traversals
Andrew Choi,
Dezhong Tong,
Brian Park,
Demetri Terzopoulos,
Jungseock Joo,
M. Khalid Jawed
IEEE Robotics and Automation Letters (RA-L), 2023
Source Code & Data
/
Video
Robust and rapid instance segmentation of DLOs via skeleton pixel traversals.
Segmentations are generated on topologically corrected skeleton structures with
intersections handled
by the simple and physically insightful optimization objective of minimizing cumulative
bending energy.
|
|
A Fully Implicit Method for Robust Frictional Contact Handling in Elastic
Rods
Dezhong
Tong*,
Andrew Choi*,
Jungseock Joo,
M. Khalid Jawed
Extreme Mechanics Letters (EML), 2023
(* equal contribution)
Source Code
/
Video
/
Oral Presentation
Fully implicit frictional contact method for simulation of slender elastic rods.
Enforces non-penetration and capable of reproducing sticking and sliding phenomena.
Case study for flagella bundling in viscous fluid shown.
|
|
Snap Buckling in Overhand Knots
Dezhong Tong,
Andrew Choi,
Jungseock Joo,
Andy Borum,
M. Khalid Jawed
Journal of Applied Mechanics (JAM), 2023
Source Code
/
Video
Study of the snap buckling process of overhand knots.
Discrete differential geometry based simulations and tabletop experiments are used to
explore the onset of buckling as a function of topology, geometry, and friction.
|
|
Preemptive Motion Planning for Human-to-Robot Indirect Placement Handovers
Andrew Choi,
M. Khalid Jawed,
Jungseock Joo
IEEE International Conference on Robotics and Automation (ICRA), 2022
Project Page
/
Video
Preemptive motion planning via human object placement prediction using human pose and
gaze
as input.
Rapidly outperforms traditional wait-and-react methods for human-to-robot pick-and-place
operations.
|
|
Weakly-Supervised Convolutional Neural Networks for Vessel Segmentation
in
Cerebral Angiography
Arvind Vepa,
Andrew Choi,
Noor Nakhaei,
Wonjun Lee,
et al.
IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), 2022
Source Code & Data
Weak supervision approaches for automated vessel segmentation.
Utilizes active contour models to generate weak labels and introduces low-cost
human-in-the-loop strategies that drastically reduce labelling cost while maintaining
segmentation accuracy.
|
|
Implicit Contact Model for Discrete Elastic Rods in Knot Tying
Andrew Choi,
Dezhong Tong,
M. Khalid Jawed,
Jungseock Joo
Journal of Applied Mechanics (JAM), 2021
Source Code
/
Video
An implicit contact model for slender elastic rod simulation capable of enforcing
non-penetration via smooth penalty energy and edge-to-edge minimum distance formulation.
Sliding friction is handled semi-explicitly.
Extensive physical validation through knot tying case study.
|
|