🚨 New preprint on transparent, tree-adaptive grounded reasoning! We introduce Bonsai, a versatile reasoning system that generates interpretable, grounded, and uncertainty-aware reasoning traces while enabling state-of-the-art performance on text and video benchmarks.

🤔 To what extent can Large Language Models (LLMs) solve unseen tasks via In-context Learning (ICL) ? 💥 We propose a new general framework ✨ ICL ciphers ✨ to quantify ``learning'' in ICL via substitution ciphers. 🔗 : Check our new preprint! arxiv.org/abs/2504.19395

5 of the 8 interdisciplinary research teams selected by the JHU Institute for Assured Autonomy that include CS faculty (feat. Anqi (Angie) Liu & more) are now completing work supported by Challenge Grants that began in 2023. Learn more: cs.jhu.edu/news/cs-facult…

Happy to share this new work! Our AI auditor needs to be adaptive (quickly) as the world is always changing and your AI model may or may not be able to maintain performance.

Welcome!

Welcome!

Welcome!

We customized uncertainty quantification and risk control techniques to help predict the safety level of cancer treatment plans, hoping to reduce burdens of clinicians without sacrificing safety. A long way to go, but a great starting point to collaborate with medical physicists!

There have been various efforts on disentangling "task learning" vs "task recall" in LLMs. We've recently explored a fresh angle by borrowing from cryptography: with substitution ciphers, we transform a given task into an equivalent, but cryptic (no pun intended!!) forms.

I am pleased to announce that I will be joining Johns Hopkins University's Computer Science Department JHU Computer Science as an Assistant Professor in Fall 2025. I am grateful to my mentors for their unwavering support and to my exceptional PhD students for advancing our lab's vision.

Welcome!

Welcome!

Welcome!

Welcome!

I am not going to make this ICML but Drew Prinster (same @ other app) will be there! Check out our recent work on online AI monitoring under a non-stationary environment.

How can the accuracy and accessibility of a leading splice-site prediction tool be improved for broader genomic research across different species?eLife - the journal @johnhopkinsuni "OpenSpliceAI: An efficient, modular implementation of SpliceAI enabling easy retraining on non-human species"

check out our new eLife - the journal paper, led by Kuan-Hao Chao along with Alan Mao, Anqi (Angie) Liu, Mihaela Pertea: OpenSpliceAI: An efficient, modular implementation of SpliceAI doi.org/10.7554/eLife.…

JHU researchers including Ziang Xiao and Anqi (Angie) Liu have created a system that could make social robots more effective at detecting and managing user interruptions in real time—a breakthrough for areas like health care and education where natural conversation is crucial.

Our university news wrote a nice story about our exhibit at the Smithsonian National Museum of Natural History. It was such a great experience, with lots of interesting and insightful conversations with visitors from all over the world! JHU Computer Science hub.jhu.edu/2025/08/05/com…

How to estimate an ML model’s performance in the wild? Our takeaway: if you can tell the difference between a target distribution and source one, it will help! Sounds natural?It turns out no previous theory directly indicates it and no previous methods exist. Check Aayush’s work!