It's a pretty cool group therapy #EMNLP2024

It was so nice to reunion with Kenneth (my PhD committee member) and every #PSU students and alumni!

🚀Arthur is on the job market for TTAP position! I am always inspired by his hard-working, innovative ideas and deep expertises in NLP and HCI... Any place will be incredibly lucky to have him!

💡 Nourah Salem, a fourth-year PhD student in the Comp Bio Program at the Graduate School is using ChatGPT to identify things we know we don't know. Salem is part of the Gao Lab (yanjungao) Explore how our students are embracing AI in the classroom ▶️ bit.ly/3UdiRK0

☀️had a great #EMNLP2024 ! thanks for stopping by our poster! I really enjoyed answering the questions and discussion on LLM embeddings for EHR and clinical tasks ⚕️🏥 🧩happy to see that the hype of #LLMs cools down a bit and foundational #NLP problems are back!

Mitigating racial bias from LLMs is a lot easier than removing it from humans! Can’t believe this happened at the best AI conference NeurIPS Conference We have ethical reviews for authors, but missed it for invited speakers? 😡

I'm shocked to see racism happening in academia again, at the best AI conference NeurIPS Conference. Targeting specific ethnic groups to describe misconduct is inappropriate and unacceptable. NeurIPS Conference must take a stand. We call on Rosalind Picard Massachusetts Institute of Technology (MIT) MIT Media Lab to retract and

Explore yanjungao research on LLMs’ capabilities in understanding and representing structured electronic health record data for clinical prediction, which was accepted at the EMNLP 2025 ▶️ bit.ly/4iI6Vu8

I don't think prompting LLMs for their reasoning is truly a good idea to enhance "interpretability". They could make a story sound really "true" and impressive, but it has nothing to do with what they truly "think". There is only one way to know what they think.

The LARK Lab led by yanjungao is revolutionizing healthcare by leveraging AI and NLP to develop groundbreaking technologies. Their work focuses on solving critical healthcare challenges and creating tools that integrate into healthcare systems ▶️ bit.ly/4i8NBp4

There’s a lot of buzz about AI, but it’s important to separate fact from fiction. AI isn't here to replace doctors—it’s here to support them! yanjungao explored how LLMs can help clinicians make decisions but still need validation at the CU Anschutz Medical Campus Engaging with AI event.

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! cu.taleo.net/careersection/…

我只能说 prompt engineering还是有大用

We’re (Northeastern U.) now partnering with CU Anschutz Medical Campus to bring CareMap into real ICUs. Because access to information isn’t a bonus— It’s part of care. 💙

I'll be in #CHI2025 ACM CHI Conference My very first CHI in-person conference. Looking forward to reunion with some friends and make new friends! And it has been 7 years since my last trip to Japan and can't wait to be in Yokohama!

I still think many work oversimplifies LLM evaluation. They talk about evaluating hallucinations, factual correctness, accuracy, alignment, harm, etc. Many of these are grounded in real-world knowledge, which come down to propositions. But the rest is all about how we agree on

Check out our new preprint on analyzing retrieval head of LLMs in the RAG set up! Our findings provide a mech interp lens that helps understand how specific attention heads contribute to drawing useful context for correct outputs🔍

Doing a PhD is a life-changing experience, and having the right mentor makes all the difference. I’ve been working with Dima since my postdoc: a deep thinker, thoughtful mentor, respectful collaborator and fun friend. someone you can trust to do solid, meaningful science. Highly

I'm not at #ACL2025 in-person this year but Ruizhe is! Come and say hi to Ruizhe if you are around!

AI is summarizing patient data, but is it doing it well? yanjungao and team answered this question by building a tool to evaluate the quality of AI-generated summaries. This is a key step toward safely integrating AI into clinical settings ➡️ bit.ly/46DozvC