Shobhita Sundaram
@shobsund
Followers
968
Following
222
Statuses
80
PhD student at @MIT_CSAIL working on deep learning | previously interned at @GoogleDeepMind
Joined August 2022
Personal vision tasks–like detecting *your mug*-are hard; they’re data scarce and fine-grained. In our new paper, we show you can adapt general-purpose vision models to these tasks from just three photos! 📝: 💻: (1/n)
5
54
309
RT @ishapuri101: [1/x] can we scale small, open LMs to o1 level? Using classical probabilistic inference methods, YES! Joint @MIT_CSAIL / @…
0
55
0
RT @ChenMaggieSY: Happy to share our work out in @AAOJournal on evaluating the performance of RETFound, a foundation model developed by @pe…
0
3
0
RT @KempeLab: Submit to our New Frontiers in Associative Memories workshop @iclr_conf. New architectures & algorithms, memory-augmented LL…
0
13
0
RT @materzynska: 1/ 🎨✨ If an AI has never seen art before, can it develop artistic capabilities? Our team from MIT, Northeastern and Shangh…
0
150
0
RT @materzynska: Our project demo is LIVE! 🎨✨ Check out the Art-Free Diffusion model and the Art Adapters build on top of it! https://t.co/…
0
13
0
RT @tyleryzhu: Have you ever wondered why we don’t use multiple visual encoders for VideoLLMs? We thought the same! Excited to announce o…
0
13
0
RT @akarshkumar0101: Very excited to share ASAL! Artificial Life aims to recreate natural evolution, but is severely bottlenecked by hand-…
0
27
0
This project came out of a wonderful collaboration with @juliachae_ (my co-first-author) along with @YonglongT , @sarameghanbeery , and @phillip_isola . (n/n)
1
0
4
I'm at #NeurIPS2024 this week! Excited to present our work on perceptual alignment with @xkungfu on Thursday morning (poster #1302). Please reach out if you want to chat about synthetic data, alignment, data-centric AI, representation learning, or anything else!
What happens when models see the world as humans do? In our #NeurIPS2024 paper we show that aligning to human perceptual preferences can *improve* general-purpose representations! 📝: 🌐: 💻: (1/n)
0
7
60
RT @EdwardVendrow: 🎯 How can we empower scientific discovery in millions of nature photos? Introducing INQUIRE: A benchmark testing if AI…
0
10
0