Research Scientist @Google Cloud AI Research | Teaching AI to complete tasks that humans need but don’t want to do.
SF, CA
Joined November 2015
- Tweets88
- Following829
- Followers203
- Likes589
Arijit Ray retweeted
HLE tested knowledge. ALE tested digital work. PLE tests the work that builds the physical world.
Today we’re releasing Physical-Engineering’s Last Exam (PLE) from Google Cloud AI Research.
All frontier models scored under 50% 🧵
Arijit Ray retweeted
Very excited to announce Gemini 4 Argon, our new Frontier AI model, and a huge step forward across key capabilities. We’re focused on rolling it out responsibly starting with government and trusted cyber defenders through our Fairwind Program today, before wider availability soon
Arijit Ray retweeted
Launching my research group, MAGIC (Manipulation and General Intelligence Control) Lab @NUSComputing, Singapore!
We focus on building the next generation of human-centric models for robotic manipulation — deployable safely, reliably, and easily in the real world. Our research spans MLLM reasoning, 3D vision, robot learning, simulation, dexterous manipulation, and cross-embodiment learning.
Interested in joining? Sign up here and I'll send a reminder email: forms.gle/oJPLR2pLTt8kLCxZ7
Arijit Ray retweeted
Today we're releasing MolmoWeb, an open source agent that can navigate + complete tasks in a browser on your behalf.
Built on Molmo 2 in 4B & 8B sizes, it sets a new open-weight SOTA across four major web-agent benchmarks & even surpasses agents built on proprietary models. 🧵
Arijit Ray retweeted
🚀 Excited to share that my team at Meta just launched Segment Anything 3! SAM 3 doubles the performance of existing models on open-vocabulary instance segmentation on our new SA-Co benchmark, with 207K unique object labels. Huge congrats to the team, so proud of this work!
Today we’re excited to unveil a new generation of Segment Anything Models:
1️⃣ SAM 3 enables detecting, segmenting and tracking of objects across images and videos, now with short text phrases and exemplar prompts.
🔗 Learn more about SAM 3: go.meta.me/591040
2️⃣ SAM 3D brings the model collection into the 3rd dimension to enable precise reconstruction of 3D objects and people from a single 2D image.
🔗 Learn more about SAM 3D: go.meta.me/305985
These models offer innovative capabilities and unique tools for developers and researchers to create, experiment and uplevel media workflows.
Arijit Ray retweeted
MLLMs are great at understanding videos, but struggle with spatial reasoning—like estimating distances or tracking objects across time.
the bottleneck? getting precise 3D spatial annotations on real videos is expensive and error-prone.
introducing SIMS-V 🤖
[1/n]
Arijit Ray retweeted
This year's @techreview #TR35 Asia Pacific honors a trio of familiar faces: #UWAllen professors @simonshaoleidu & @ranjaykrishna, and PhD alum @sewon__min of @allen_ai & @Berkeley_EECS! Read about their work advancing #AI, #LLMs, computer vision and more: news.cs.washington.edu/2025/…
ALT Side-by-side portraits of Simon Shaolei Du, Ranjay Krishna and Sewon Min
Arijit Ray retweeted
“Spatial thinking is the foundation of thought, moving in spaces essential to life." - Barbara Tversky in Mind in Motion.
Spatial reasoning goes beyond just reasoning about what is around you. It's reasoning about dynamics, actions, egocentric motions, and much more.
Test your VLMs on SAT! They still can't do it.
Thrilled to announce that our paper SAT has been accepted to #COLM2025! 🎉
Better yet— all the data and code are already open-sourced. Dive in, experiment, and let us know what you build!
Data:huggingface.co/datasets/arra…
Code:github.com/arijitray1993/SAT
Arijit Ray retweeted
Thrilled to announce that our paper SAT has been accepted to #COLM2025! 🎉
Better yet— all the data and code are already open-sourced. Dive in, experiment, and let us know what you build!
Data:huggingface.co/datasets/arra…
Code:github.com/arijitray1993/SAT
Arijit Ray retweeted
How should a robot hold a water bottle? 🤔 That depends: is it opening it, or passing it to you?
I’m excited to introduce GraspMolmo, a VLM that predicts semantically appropriate grasps based on your command!
Website: abhaybd.github.io/GraspMolmo…
🧵 Thread ↓
Arijit Ray retweeted
🤔What tasks do we want robots to handle? Are these preferences based on saved time or feelings we associate with the tasks?
Introducing Why Automate This?—a study exploring automation preferences across social groups, using feelings & time-spent as key factors. 👇 (1/5)
Arijit Ray retweeted
As my kids are singing APT non-stop these days, I did a bit of reverse engineering of the APT music video and tried to understand why the MV is so addictive.
Here is what I learned.
Arijit Ray retweeted
Check out our new benchmark to test both spatial and dynamic reasoning capabilities of MLMs! 🧭🏃
🤖🌍 These two capabilities are especially important when grounding VLMs on real agents that can physically interact with the real world.
Arijit Ray retweeted
Congratulations to @petsiuk on receiving the ECCV Best Paper Honorable Mention Award!
Arijit Ray retweeted
🐨Koala: Key frame-conditioned long video-LLM
Koala is a new video-LLM that can answer questions about longer videos than previously possible.
--with R. Tan, X. Sun, P. Hu, J. Wang, H. Deilamsalehy, B.A. Plummer and B. Russell
Link to paper and demo below
Arijit Ray retweeted
just for fun, I made NeurIPS GPT -- a custom GPT that has the list of all @NeurIPSConf 2023 paper titles, abstracts and authors.
chat.openai.com/g/g-T9fiOn4c…
Arijit Ray retweeted
Introducing💥Lasagna: a layered diffusion model for image relighting. Lasagna adds realistic lighting to input images, even to vector art!
Joint work with @drbashkirova @ARRay693 Rupayan Mallick, @SarahAdelBargal @RanjayKrishna @jianming_zhang_
arxiv.org/abs/2312.00833
Arijit Ray retweeted
You know which one will fall last? Robotics.
I predict an AI will discover important new theorems BEFORE we have a generally-capable robot that can cook, clean, drive, patrol, and take care of your pets, babies, sick, and elders in every household.
We humans assume a professional mathematician is way harder than manual labor. Our AI systems will disagree.
Don’t ever under-estimate the Moravec’s paradox.
