Thanks Kartheek!

Thanks for the recognition @CVPR 🙏

Check out this amazing work by @Harman26Singh and team @AIatMeta! Traditionally, agents are evaluated on how good they are at a task. But as agents run longer, an equally important question is how efficiently they improve with experience (plasticity). Turns out capability and plasticity are pretty different things.

Congrats, Pragya!!

This is amazing! Congrats!!

Super thrilled to share that our work, SRL-CLIP, won the Best Paper Award at the CV4Smalls #CVPR2026 workshop! 🏆 Huge thanks to my amazing co-authors @zeeshank95 and @MakarandTapaswi . Also, a big thank you to the organizers (@shaydamoezzi and team) for putting together such a thoughtful workshop. With the current trend of "scaling everything," it is incredibly refreshing to focus on data inefficiency as a design opportunity rather than a limitation. Data efficiency is one of the main pillars needed to improve real-world video understanding.

MINERVA-Cultural is a step toward helping the world build more equitable and diverse VLMs 🌍. Huge thanks to all my co-authors and the team!. Check out the paper here: arxiv.org/abs/2601.10649 and the dataset here: github.com/google-deepmind/n… #CVPR2026 #AI #ComputerVision (7/7)

This allows us to independently assess each intermediate step of the model's thought process. 🔍 Through our graph-based error tracing, we discovered that roughly 75% of all model failures actually stem from the visual perception of cultural elements. 🤯(6/7)

Vanilla accuracy just tells us IF a model failed, but not WHY. To address this, we propose a novel multi-turn algorithm for fine-grained error analysis: ✓ Leverages human reasoning traces ✓ Builds formal "evidence graphs" ✓ Compares graphs directly against the model's thinking traces (5/7)

These tasks demand deep multimodal reasoning across six cultural domains (Sports, Cuisine, Festivals, Tourism, Rituals, and Education) on videos ranging from one minute to over an hour in length! The performance gap is massive: human performance sits at 95%, while the best SoTA model reaches only 50%. (4/7)

To ensure true cultural authenticity, we didn't rely on synthetic data. Instead, we built a rigorous, multi-stage human curation pipeline: ✓ Human Curators crafting complex questions & reasoning traces ✓ Independent Auditors answering blindly to validate objectivity ✓ Continuous, dialogue-driven feedback loops between both groups to resolve ambiguity and guarantee high quality (3/7)

We introduce MINERVA-Cultural: a benchmark for multicultural + multilingual long-video QA across 18 global locales. ✓ 540 videos, 2,400 questions ✓ Native-language questions (no machine translation) ✓ Human-written multi-step reasoning traces ✓ Evidence-graph based error diagnosis Humans: 95.22% | Best model: 50.33% (2/7)

Frontier models have become excellent at understanding videos. But what happens when we test them outside the comfort zone of Western, English-centric data? In our #CVPR2026 (Highlight) work, we pushed these models to their limits to see if they can function effectively in diverse global contexts. The results? They are struggling. Work done with @NagraniArsha @skawshik11 @Harman26Singh @dinesh_tewari1 @0xtob @CordeliaSchmid Anelia Angelova @shachi_dave (1/7)

Hi all! Today I will present our work, SRL-CLIP, a recipe to efficiently adapt CLIP to videos at the CV4Smalls CVPR workshop in Room 102. I have an oral talk scheduled at 2 pm followed by the poster session. Please do stop by! Work done with @zeeshank95 @MakarandTapaswi #CVPR2026

Grateful to be recognized as an Outstanding reviewer at #CVPR2026. Always happy to do my part and give back to the community. This CVPR is very special for me because we also have 1 Main paper (Highlight) and 1 workshop paper (Best paper candidate). See you all in Denver! 🚀