Suyu (Michael) Ye

@suyu_ye

CS PhD @JohnsHopkins

suyuye.com

Joined August 2022

106 Following

53 Followers

Replies Reposts Media

Suyu (Michael) Ye
@suyu_ye
Oct 7
Really cool work!
Homanga Bharadhwaj
@mangahomanga
Oct 7
NEVER LOOK BACK! Introducing LEDGER: persistent 3D object memory from egocentric videos. Remember where things were, when they moved, and what they held -> then answer questions without looking back at the video! ledger-3d.github.io
Suyu (Michael) Ye
@suyu_ye
Oct 5
What can you learn about a robot partner that still matters when the task changes? Its physical limits. We introduce the problem of inferring those limits by watching it work with another robot, then coordinating with it zero-shot on a new task. Introducing Watch, Infer,
1
Suyu (Michael) Ye
@suyu_ye
Oct 2
The world changes. World models should too. Check out our newest work!
Brian Zheyuan Zhang
@zheyuanzhang99
Oct 2
A pretrained latent world model's predictions can become unreliable when test-time dynamics differ from training. Introducing JEPA-TTT, which enables persistent test-time training of latent world models for planning under dynamics shifts. Project Page: jepa-ttt.github.io
Suyu (Michael) Ye
@suyu_ye
Jun 1
Excited to share our newest work!
Yifan Yin
@yifanyin_11
Jun 1
What should a world model capture for embodied agents? An agent acting under partial observability needs a belief over the 3D world: what it has seen, what may exist beyond view, and how that belief should update as it moves. Introducing 3D-Belief, a 3D world model for embodied
Suyu (Michael) Ye
@suyu_ye
Mar 9
🚀 New robot coordination framework How should robots communicate to coordinate safely in shared environments? Directly mapping language to robot motion is risky and hard to verify. Our work proposes a different approach. Introducing CaPE: Code as Path Editor. 🧵[1/6]
2

Suyu (Michael) Ye

Suyu (Michael) Ye’s posts