📢'The World Is Your Oyster' is for those who are born in global north countries--don't experience visa bureaucracy & emotional burden that non-white global south scholars do. Citizenship is a privilege that we need to recognize in the academe. 1/n

nature.com/articles/d4158…

[1/7] 🚀 Introducing the Language Model Transparency Tool - an open-source interactive toolkit for analyzing Transformer-based language models.

We can't wait to see how the community will use this tool!
github.com/facebookresear…

I was so disappointed to see an episode of Abbott Elementary which assumed that tools purporting to detect 'AI' generated text work 100%, given the number of students wrongfully accused of using ChatGPT because of these tools.

Summary table of existing LLM unlearning methods/problems. Some interesting discussions on associated challenges too, precisely defining the unlearning scope, lack of generality in current solutions beyond specific tasks/contexts, and (ofc) lack of evals. arxiv.org/abs/2402.08787

How to get paid is increasingly the biggest puzzle facing biomedical #AI -- more so than developing the technology. Our new NEJM AI article discusses the economics of medical AI + payment landscape ai.nejm.org/stoken/default…

Real pleasure writing this w/ Michael D Abramoff and…

We are counting down to the annual ARVO meeting! From the latest risk factors and treatments for cataracts and AMD to deep learning and ophthalmic imaging, we hope you will join our department for our vast array of presentations and sessions this year! #ARVO2024 #CUAnschutz

🚨BREAKING: The Stanford University Institute for Human-Centered AI publishes its Artificial Intelligence Index Report 2024, one of the most authoritative sources for data and insights on AI. Below are its top 10 takeaways:

1. AI beats humans on some tasks, but not on all;
2. Industry…

The way how its going, might start expecting grants in future apps.

👇👇👇

Looks super cool... Haven't read the work, but watching the video, seems an interesting use of super-resolution technique for OCTs.

A novel #AI algorithm applied to an adaptive optics OCT system produced retinal images 100 times faster and with 3.5x improved image contrast.

Video by Vineeta Das (NEI Intramural program) about her and group’s Communications Medicine paper: youtu.be/TYtU_I0DHS4

1/2

😂😂😂

Happy to announce our foundation model for wearables published at npj Digital Medicine today. This model sets a new standard in #Wearables , significantly outperforming human activity recognition benchmarks in diverse conditions. 🚀

nature.com/articles/s4174…

Microsoft presents Rho-1

Not All Tokens Are What You Need

Previous language model pre-training methods have uniformly applied a next-token prediction loss to all training tokens. Challenging this norm, we posit that 'Not all tokens in a corpus are equally important for

eye doctors after the solar eclipse

🧵Introducing the 'PromptWISE' (Prompt engineering for Well-structured, Interactive, and Supportive Education) paradigm

Our study focuses on evaluating the impact of the PromptWISE paradigm, aiming to enhance interactions between patients and #LLMs for medical education. [1/9]

Our 12 scaling laws (for LLM knowledge capacity) are out: arxiv.org/abs/2404.05405. Took me 4mos to submit 50,000 jobs; took Meta 1mo for legal review; FAIR sponsored 4,200,000 GPU hrs. Hope this is a new direction to study scaling laws + help practitioners make informed decisions