Sourajit Saha
424 posts
@sourajitCS
Computer Vision | Multimodal and Interactive Search and Retrieval | Video Understanding | PhD Student @ UMBC, Ex intern@JHU
🎉 Excited to share that our paper, PARSE, has been accepted to #ECCV2026! Can a text-to-image model forget one unwanted concept without forgetting everything around it? PARSE is our training-free approach to doing exactly that. More below 🧵
We are grateful to all of the 17,491 reviewers who helped make #CVPR2026 possible. We are especially pleased to recognize the following Outstanding Reviewers, whose high-quality reviews (as judged by their Area Chairs) placed them among the top 5% of reviewers.
🚀MIT Flow Matching and Diffusion Lecture 2026 Released (diffusion.csail.mit.edu)! We just released our new MIT 2026 course on flow matching and diffusion models! We teach the full stack of modern AI image, video, protein generators - theory and practice. We include: 📺 Videos: Step-by-step derivations. 📝 Notes: Mathematically self-contained lecture notes 💻 Coding: Hands-on exercises for every component We fully improved last years’ iteration and added new topics: latent spaces, diffusion transformers, building language models with discrete diffusion models. Everything is available here: diffusion.csail.mit.edu A huge thanks to Tommi Jaakkola for his support in making this class possible and Ashay Athalye (MIT SOUL) for the incredible production! Was fun to do this with @RShprints! #MachineLearning #GenerativeAI #MIT #DiffusionModels #AI
Excited to share our new work: Generative Video Motion Editing with 3D Point Tracks. We propose a framework that uses 3D point tracks to precisely edit both camera and object motion in a video, unlocking a wide range of new editing applications.
Perona and Belongie wondered if they could replace search words with images, but they knew that it would not be easy to convince people to tediously and appropriately tag every object and its parts in millions of pictures. caltech.edu/about/news/pie…
Meta Superintelligence Labs presents MetaEmbed: Scalable multimodal retrieval • Flexible late interaction via Meta Tokens • Test-time scaling: trade off retrieval accuracy vs efficiency • SOTA on MMEB + ViDoRe, robust up to 32B models • Matryoshka training → coarse-to-fine multi-vector embeddings
🚨 Our paper “Side Effects of Erasing Concepts from Diffusion Models” has been accepted to EMNLP 2025 (Findings)! #EMNLP2025 We investigate the vulnerabilities of Concept Erasure Techniques (CETs) Big shoutout to my amazing collaborators @sourajitCS @manasgaur90 @trgokhale 1/n
