Embedded Files
Journal Club
Journal Club
Image-to-text embeddings (Praful, we can do it next week, with help from gemini ;-)):
Image-to-text embeddings (Praful, we can do it next week, with help from gemini ;-)):
A visual–language foundation model for pathology image analysis using medical Twitter
Zhi Huang, Federico Bianchi, Mert Yuksekgonul, Thomas J. Montine & James Zou
Nature Medicine volume 29, pages2307–2316 (2023)Cite this article
______
Related articles
A visual-language foundation model for computational pathology
https://www.nature.com/articles/s41591-024-02856-4
https://www.nature.com/articles/s41379-020-0540-1
https://www.nature.com/articles/s41467-024-50043-3
______
IARC
IARC
Example of data api call: https://gco.iarc.fr/gateway_prod/api/overtime/v2/20//data/population/0/2/752_356/(16)/?ages_group=0_17&years=2017&year_start=2017&year_end=2017&cumrisk_trend=1
Reference Info
Reference Info
https://ci5.iarc.fr/ci5plus/download
https://ci5.iarc.fr
https://gco.iarc.fr/media/ci5/data/ci5plus/CI5plus_Detailed.zip ?
Hackathon
Hackathon
Activate gemini nano in your browser: https://medium.com/google-cloud/get-started-with-chrome-built-in-ai-access-gemini-nano-model-locally-11bacf235514
Once again, the Web is where AI happens: it turns out you'll have nano native in chrome. The intersection between AI and Web computing is where it happens :-P !
Thank you Jeya for keeping track of the news :-D
https://www.youtube.com/watch?v=VDtgyEB9Q_Q&t=4s
https://docs.google.com/document/d/1Bvd6cU9VIEb7kHTAOCtmmHNAYlIZdeNmV7Oy-2CtimA/edit
Google Chrome Built-in AI - Early Preview Program
Thanks for your interest in Chrome's built-in AI! We're excited to have you on board.
While we'll be adding those interested to the Early Preview Program (EPP) soon, there's *no need to wait to start exploring*! You can jump right into exploring the APIs, prototyping ideas, and joining the conversation today.
Get started right away:
- Setup and How-Tos for all our APIs: https://goo.gle/chrome-ai-dev-preview-index
- Discussions: https://goo.gle/chrome-ai-dev-preview-discuss
- Take on the Built-in AI Challenge: https://goo.gle/ChromeAIChallenge
*What about the EPP?*
The EPP provides valuable benefits like early access to information about upcoming changes, sneak peek of new APIs, and a direct line to share your feedback and shape the roadmap.
Let's unleash the power of AI on the web together - one inference at a time! 🚀
Chat away with nano
10:30
10:30
Leveraging Advanced Computer Vision for Medical Imaging Applications
BIOSTATISTICS BRANCH SEMINAR SERIES PRESENTS
BB seminar:
Cornelia Fermüller, Ph.D.
Research Scientist
University of Maryland Institute for Advanced Computer Studies (UMIACS) & Maryland Institute for Health Computing
Date: Wednesday, November 20h, 2024
Time: 10:30 am to 11:30 am (EST.)
Location: 4E032/034 & Join Via WebEx
Meeting number: 2313 981 7402
Password: BBsem11.20
Join by phone: 1-650-479-3207 Call-in toll number (US/Canada)
Access code: 2313 981 7402
Title: Leveraging Advanced Computer Vision for Medical Imaging Applications
Abstract:
In this talk, I will explore three applications of cutting-edge computer vision techniques for medical imaging and discuss the unique challenges associated with each. First, I will cover a semi-supervised training approach designed to address the scarcity of precise annotations in chest CT scans by incorporating weakly labeled data into deep learning algorithms. Next, I will introduce innovative methods for aligning and analyzing highly noisy sonogram images, advancing the feasibility of a new device for sleep apnea detection. Finally, I will discuss biomarker identification techniques for tracking progression in multiple sclerosis using MRI scans, where data scarcity and label inconsistency pose significant hurdles. I will also provide an overview of the upcoming initiatives at the newly established Maryland Institute for Health Computing, which I recently joined, and how we aim to drive future advancements in health-focused AI.
Report abuse