←
Jadelynn Dao8 posts
Jadelynn Dao avatar

Jadelynn Dao

@_jadelynn

CS PhD @Stanford @StanfordAILab | @NVIDIAAI

◷ Joined October 2024▧ 1 Photos and videos
Emily Jin avatar
⟳ Jadelynn Dao retweeted
Emily Jin@emilyzjin·Jul 10

During long-horizon task planning, a robot must decide what to do, while ensuring each action is geometrically feasible. To this end, we propose 𝗽𝗹𝗮𝗻𝗻𝗶𝗻𝗴 𝘄𝗶𝘁𝗵 𝗶𝗻𝘁𝗲𝗿𝗹𝗲𝗮𝘃𝗲𝗱 𝘃𝗶𝘀𝗶𝗼𝗻-𝗹𝗮𝗻𝗴𝘂𝗮𝗴𝗲 𝘁𝗵𝗼𝘂𝗴𝗵𝘁𝘀.

0:17
Milan Ganai avatar
⟳ Jadelynn Dao retweeted
Milan Ganai@milanganai·Jun 30

🤔 Can VLA models teach themselves the best way to reason? We introduce R&B-EnCoRe which is a self-improving pretraining cycle for reasoning VLAs to propose and self-discover what embodiment-specific reasoning looks like. 🔗 milanganai.github.io/rnb-enc… 📃 arxiv.org/pdf/2602.08167

1:02
Jadelynn Dao avatar
Jadelynn Dao@_jadelynn·Jun 11

4/4 we are incredibly excited about this line of work!! feel free to reach out to us 😁 📄 arxiv.org/abs/2606.12402 🔗 jadee-dao.github.io/direct/ 🤖 jadee-dao.github.io/direct/v…

Jadelynn Dao avatar
Jadelynn Dao@_jadelynn·Jun 11

3/4 so, when and where is test-time compute in robotics worth its cost? DIRECT 🧭 (Dynamic Inference Router for Embodied Compute Tradeoffs) is a lightweight router that reads scene + instruction context and sends each task to the cheapest planner that can still solve it

Jadelynn Dao avatar
Jadelynn Dao@_jadelynn·Jun 11

2/4 vlms are increasingly deployed as high-level planners for robots, and the prevailing strategy has been to scale test-time compute to boost capability but more thinking, bigger models, and longer memory all add latency, tokens, and FLOPs, with diminishing, uneven returns