Wei-Hung Weng
@ckbjimmy
Followers
47
Following
33
Statuses
3
Research Scientist @GoogleAI working on Medical AI, MD/PhD, ex @MIT_CSAIL @HarvardDBMI
Joined April 2024
Pleased to announce the release of the open-source, fully re-labeled MedQA benchmark, to facilitate more accurate model performance estimation. GitHub repository: Check David's post for more details!
Part of our Med-Gemini work was a full relabelling of MedQA, revealing that at least 7.4% of examples are unfit for evaluation. Today, we open sourced these annotations alongside our evaluation script as a new standard evaluation on MedQA. A thread 🧵:
0
1
7
Excited to share the work using conformal abstention to abstain LLM hallucinations!
Happy to share our latest work using conformal abstention to improve factuality of LLMs by avoiding hallucinations: A short thread 🧵:
0
0
4
Excited to share Med-Gemini! Except for remarkable SoTA performance on MedQA and various medical language/image/video tasks, one super exciting potential to highlight is to unlock novel applications via the long-context capability and multimodal dialogue!
Delighted to share ✨Med-Gemini✨ - our new family of multimodal models for medicine unlocking new possibilities for health - More accurate multimodal conversations about medical images🩻, surgical videos📽️, genomics🧬, ultra-long health records📚, ECGs🫀 & more with state-of-art performance across multiple benchmarks More accurate, up-to-date answers to medical questions with advanced reasoning and intelligent use of web-search Long-context abilities. Summaries or referral letters from long health records, analyses of dozens of long research PDFs & more (1/6)
0
1
18