The Students Offering Support (SOS) Program aims to assist students from disadvantaged backgrounds (including, but not limited to, first generation students, students from low income households, etc.) as they apply to MAS. A graduate student volunteer will provide one round of meaningful feedback on either the applicant's portfolio or statement of objectives. Participation in SOS does not guarantee admission. Unfortunately, international applicants who are studying or employed outside of the US are not eligible to participate in the SOS Program. Apply by Nov. 1, 2025.
https://t.co/zg24UUV3I1
Tired of seeing top robotics papers get 3 citations while fluff gets 300?
We fixed that. Our sound‑localization work just dropped and the media already gave it a better title:
“Tsinghua and Michigan team cuts annotation cost from millions to ZERO.”
That’s not hype. It’s real.
No labels. No human in the loop. Ego‑motion + geometry = free supervision. Works on real robots. Open‑source.
If you do navigation, manipulation, or multimodal perception – this is your new baseline.
If it saves you months of labeling, share it.
🔗 Paper + Code in reply 👇
Spent 6 months on a CVPR highlight Top1% paper. 3 citations.
Saw a simpler version of a similar idea win CORL Best Paper.
Meanwhile, a demo-version video2audio demo I built in one day during my internship at Pika hit 200k+ views.
Not frustrated — just genuinely curious: how do papers actually get recognized? Is it about novelty, timing, visibility, or something else?
Would appreciate any honest thoughts from folks who've been through this. #Research#CORL#Pika#AcademicLife#MachineLearning#AI
ChatGPT often feels “warm,” subtly encouraging self-centered narratives. DeepSeek and Claude, by contrast, feel more neutral. It raises an interesting question: how much does model personality shape our personality during use?
Doing research in a time when every other week someone launches an AI startup and big companies move 1000× faster. Feels like standing still in a storm.
I've deleted this tweet so many times trying to figure out how to connect previous research and internship work with current trends to draw attention. Getting used to tweeting in public without overthinking is a real cultural gap every nonnative researcher has to adapt to.