🚀 Excited to share our new paper! We show how instruction templates hugely impact Multimodal Language Models performance and introduce a template generator for robust evaluation & instruction tuning. Paper: arxiv.org/abs/2412.08307 Code: github.com/shijian2001/Te… 👇for more...

Excited to share that I’m joining Adobe as an Applied Scientist on the GenAI team led by Yunyao Li! Looking forward to learning, collaborating, and enhancing the user experience of our AI assistant of @AdobeExpCloud

🚨Deadline Extension Alert! 📅We’ve extended the direct submission deadline to March 30, 2025! Submit your work on NLP for positive impact, social good and real-world applications. 🌍Special theme: NLP for Climate Change. 📢More details: sites.google.com/view/nlp4posit… #NLP4PI #ACL2025

🌟Check out our new paper on geographic bias in VLMs!

Want to know what your LLM don’t know? This is how 👇 Preprint: arxiv.org/abs/2503.23361 Code: github.com/uscnlp-lime/SEA

What’s left for agentic memory research then😲

🚀 #CHI2025 Struggling with syncing, brainstorming, and collaborating in co-located group discussions? ✨We've designed LADICA to provide in-person teams with powerful meta- and macro-cognitive support while working in a shared digital workspace (1/n)

I've always felt that LLMs fall short when it comes to multi-agent problems. What we truly need is a world model—one that can infer intent, beliefs, and theory of mind. If you're at ICLR, make sure to check out Hongxin's poster! As a bonus, he might give you a sneak peek at

Reasoning about the "why" behind user behavior can improve LLM personas! ✨🧠📈 📝Excited to share our new work: Improving LLM Personas via Rationalization with Psychological Scaffolds 🔗 arxiv.org/abs/2504.17993 🧵 (1/n)

First day in Microsoft! Excited for the summer in Seattle🤩

Our paper just got accepted by COLM2025! See you at Montreal for further discussion on how to discover LLM deficiencies! #colm2025

Thank you so much Caiming! We show that involving coding as a new type of action apart from GUI action for CUA can significantly help improve the computer-using performance while reducing the total actions for task solving. If you are interested in it, please take a look at