LARK NLP Lab

Language, Reasoning and Knowledge (LARK) Lab at @CUBiomedInfo . We focus on NLP, #AI4Health, #AI4Science

[yanjungao @Serena_pancakes]

Back from #ACL2026. One thing that stayed with me: while AI is racing toward bigger models and new systems, several ACL keynotes revisited language, people, and intelligence. Years later, after spending my days working on LLMs, I realized I’m still fascinated by the same questions.

A big shout out to Dr. He Cheng, postdoc in @lark_nlp_lab presenting the #LogosKG paper (https://t.co/hQvNjQ2cj4)!

[CU Department of Biomedical Informatics (DBMI) @CUBiomedInfo]

Resarchers in the @lark_nlp_lab at @CUAnschutz developed MUSE, one of the first multi-LLM uncertainty algorithms grounded in information-theoretic principles.

Instead of relying on a single model’s output, MUSE leverages a subset of models. Learn more ➡️ https://t.co/exkKS6l9rA

[LARK NLP Lab @lark_nlp_lab]

We are hiring for #NLP #postdoc !!

[yanjungao @Serena_pancakes]

New year, new opportunities!🎉
Our lab @lark_nlp_lab is growing so fast and we have a postdoc opening focused on foundational LLM & NLP research (reasoning, uncertainty, efficiency) at CU Anschutz, recently ranked among the top 5 research hospitals globally (by Nature index).
If interested, feel free to apply: https://t.co/4aysvBdmZU

[yanjungao @Serena_pancakes]

Heading to #NeurIPS2025 San Diego CA🌴☀️!!
Our lab @lark_nlp_lab got one poster from our student and postdoc about medical KG+RL at the #GenAI4Health workshop !
I am also presenting a paper led by @liruizhe94 at the #MechInterp workshop!
Feel free to DM for ☕️
#NLP #healthAI #interpretability

[yanjungao @Serena_pancakes]

Postdoc Dr. He Cheng from our lab @lark_nlp_lab will present two work at #AMIA2025!
One will be an oral presentation at the NLP Pre-symposium workshop, on the topic of efficient knowledge graphs! (Nov 16)
And another poster presentation on the 17th on evaluating LLMs on longitudinal patient summarization.
Please come!

[yanjungao @Serena_pancakes]

Excited to present our lab @lark_nlp_lab #EMNLP2025 paper on longitudinal clinical summarization!
We test LLMs on public EHR data (MIMIC), but under a more realistic, multi-day setup, capturing evolving patient trajectories and causal links across time.
A step toward faithful, time-aware clinical AI. 🩺🧠
Paper: https://t.co/l5wYiKG6PD

[yanjungao @Serena_pancakes]

Happening now! #EMNLP2025
Our work MUSE (Multi-LLM Uncertainty via Subset Ensembles) @lark_nlp_lab
We proposed a novel, information-theoretic method that quantifies both intra- and inter-LLM uncertainty.
Paper: https://t.co/INh561IWdC

[LARK NLP Lab @lark_nlp_lab]

🥳Happy to share that, while we are celebrating the one-year anniversary, our lab has two #EMNLP2025 paper accepted (1 Main + 1 Findings)!
Thanks all the lab members and collaborators for the great work!!

[yanjungao @Serena_pancakes]

My lab @lark_nlp_lab has postdoc openings: if you are interested in developing advanced methods to enhance LLMs for healthcare applications, please apply! We have exciting projects awaiting! https://t.co/0DlwuvLWLY

[LARK NLP Lab @lark_nlp_lab]

Maya and Shiyue are presenting their paper on long document summarization in Wed Boulder NLP meeting!
Preprint available: https://t.co/2ruKXXLSB6

[LARK NLP Lab @lark_nlp_lab]

Congrats, Nourah, for passing her PhD comprehensive exam (co-advise with Prof Larry Hunter)! 🎉

Here's to the first PhD candidate at LARK Lab!