HTTP/2 200
server: GitHub.com
content-type: text/html; charset=utf-8
last-modified: Sun, 01 Dec 2024 22:58:41 GMT
access-control-allow-origin: *
strict-transport-security: max-age=31556952
etag: W/"674cea21-1f92"
expires: Mon, 29 Dec 2025 19:12:33 GMT
cache-control: max-age=600
content-encoding: gzip
x-proxy-cache: MISS
x-github-request-id: 4429:2B0FD4:93E0DF:A5D22E:6952D049
accept-ranges: bytes
age: 0
date: Mon, 29 Dec 2025 19:02:33 GMT
via: 1.1 varnish
x-served-by: cache-bom-vanm7210095-BOM
x-cache: MISS
x-cache-hits: 0
x-timer: S1767034953.166053,VS0,VE224
vary: Accept-Encoding
x-fastly-request-id: 993dfca71b527b45a47b3dd537d925cdec4c125a
content-length: 2261
Eric Cai
Eric Cai
Hello! I am a second-year Master's student in the Robotics Institute at Carnegie Mellon University, advised by Prof. David Held. I am broadly interested
in well-structured visual representations for generalizable manipulation. My ultimate goal is to develop a unified learning framework
for visual and physical intelligence.
Prior to joining Carnegie Mellon University, I spent two years as an applied data scientist at The Boeing Company, where I worked with the US Navy to develop
machine learning systems for the F/A-18 program.
TaskSeg: Unsupervised Deep Instruction Tuning for Few Shot Object Segmentation
Jessica Brown*, Teeratham Vitchutripop*, Eric Cai, Jenny Wang, David Held
Under Review, 2025
Project Page
Leveraging optical flow from demonstration videos for unsupervised novel object segmentation.
Non-rigid Relative Placement through 3D Dense Diffusion Eric Cai, Octavian Donca, Ben Eisner, David Held
8th Conference on Robot Learning, 2024
Project Page
/
Bibtex
/
arXiv
/
Code
A point cloud diffusion approach for non-rigid goal prediction, allowing for a generalizable deformable manipulation policy.
Sequential Object-Centric Relative Placement Prediction for Long-horizon Imitation Learning
Ben Eisner, Eric Cai, Octavian Donca, Teeratham Vitchutripop, David Held
Learning Effective Abstractions for Planning (LEAP) Workshop @ CoRL, 2024
Project Page
An approach modeling precise object-object relationships for long-horizon sequential manipulation tasks.
