Excited to announce the HITI Lab's release of the EMory BrEast Imaging Dataset (EMBED) on AWS Open Data containing 20,000 women and over 500,000 mammographic images. Judy Gichoya Imon Banerjee @EmoryRadiology #breastcancer #AI registry.opendata.aws/emory-breast-i…

Superstar Judy Gichoya for her amazing service at NIH. Honored to know you!! <3

Much has been made of the declining life expectancy in the US. The reality is regional. Think of the US as two countries: Some blue counties with increasing life-expectancy better than 🇯🇵 (84) & many red counties with plummeting life expectancy worse than 🇷🇺 (71). 1/

Should we train our own large language model (LLM) on domain-specific data from scratch? Researchers at Bloomberg did just that and shared a detailed technical report describing the dataset, model configuration, and training procedure. 1/9

Don’t mess with otters.. 😂

New systematic review and guidelines on self-supervised learning for medical imaging just dropped! nature.com/articles/s4174… Congrats to mars huang Akshay Chaudhari @syeung10 Anuj Pareek

NEW Comment by Hanzhou Li et al: 'Ethics of large language models in medicine and medical research'. #ChatGPT #ethics @EmoryRadiology @IUPUI Massachusetts Institute of Technology (MIT) bit.ly/3ABiUVN

Excellent TED talk from Sal Khan: - many inspiring examples of GPTs finetuned into socratic tutors, assisting without giving away answers. - none of it "out of the box", requires prompt engineering, finetuning, data collection, iteration. - sense of barely scratching the surface.

Integrating LLMs in medicine raises ethical concerns about bias, trust, authorship, equitability, and privacy. Establishing guidelines for responsible and effective use of LLMs is crucial. John T Moon Saptarshi Purkayastha @LeoAnthonyCeli Hari Trivedi Judy Gichoya @EmoryRadiology #AIethics

Our paper The Lancet Digital Health on LLM ethical concerns (bias, trust, equitability, privacy, authorship) led by Hanssen Li, MD and #AI leaders Judy Gichoya Hari Trivedi Saptarshi Purkayastha @LeoAnthonyCeli #ai #AIethics #LLMs

How to meet with your advisors/mentors? If you are a grad student, having effective regular meetings with advisors or mentors is absolutely crucial for your success! Here are some tips on how to make the most of it! 🧵

Always observe significant drop in external validation? Our anomaly detection model that supports the external validation without data sharing and maintain performance. Judy Gichoya Saptarshi Purkayastha Mayo Clinic MI2 Mayo Radiology Arizona HITI Lab Hari Trivedi ieeexplore.ieee.org/document/10122…

We present SpQR, which allows lossless LLM inference at 4.75 bits with a 15% speedup. You can run a 33B LLM on a single 24GB GPU fully lossless. SpQR works by isolating sensitive weights with higher precision and roughly doubles improvements from GPTQ: arxiv.org/abs/2306.03078🧵

Our new review in #bias in #radiology Judy Gichoya Hari Trivedi Bhavik Patel, MD MBA Mayo Clinic MI2 @EmoryRadiology Mayo Radiology Arizona sciencedirect.com/science/articl…

Couldn’t ask for a better work spouse than Judy Gichoya at the Emory AI Bias Datathon HITI Lab @EmoryRadiology

I had a blast at the #AIBiasDatathon! It was so much fun to work with this massive data, see perspectives from all around the world and catchup with friends. A big shout-out to Judy Gichoya Hari Trivedi, Janice Newsome MD FSIR and the whole team for making this possible. "=" > "÷"

#Embeddings are a way to compress image information, but are they on par with original images? Do they encode protected information? This was our teams project! Stay tuned for more amazing things from these folks! Good reading source: doi.org/10.1148/radiol… #AIBiasDatathon

Great to have been back to #AMIA2023 and reconnect with my teachers , friends and new collaborators (whom we have been working virtually with for years) as well as the very expansive Emory School of Medicine and CHOA informatics faculty.

Thrilled to share our preprint on use of #SyntheticData in medical imaging research! 📝 arxiv.org/abs/2311.09402 💬 Here is a custom GPT that will answer all your questions about the paper: chat.openai.com/g/g-iCGh2R2QA-… Some key takeaways in a 🧵 /1