🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

Excited to have finished in the top 10 in this year's RSNA The American Society of Neuroradiology American Society of Spine Radiology (ASSR) C-spine Fracture Detection AI Challenge on Kaggle! Congrats to all of the other winners & competitors, and big thanks to the organizers and annotators. Solution summary here: tinyurl.com/4zayhwnp

Who else is surprised that since 2016, no radiologists lost their job to AI ? A bit ironic that radiologists now wish to hire AI researchers from FANG labs who unfortunately lost their jobs. Btw, working with Judy Gichoya is an incredible opportunity ! #WileECoyoteVsGHinton

Now convinced to focus on teaching abstraction for my kids. It will always be useful for them in the future world where humans compete against AI. #ChatGPT

Unleashing the Power of Deep Learning for Breast Cancer Detection through Open Mammography Datasets doi.org/10.1148/ryai.2… Alexandre Cadrin-Chênevert Université Laval #BreastRad #BreastCancer #ADMANI #SAR2023 #AbdRad

Unleashing the Power of Deep Learning for Breast Cancer Detection through Open Mammography Datasets doi.org/10.1148/ryai.2… Alexandre Cadrin-Chênevert Université Laval #BreastCancer #ML #MachineLearning

Impressive but imperfect zero-shot generalization in medical imaging for Segment Anything Model (SAM) from FAIR segment-anything.com

I estimate that training a foundational vision model inspired by GPT4 or PaLM 2 models but with 2D or 3D medical images using 4T samples would need 250000 to 1000000 more computing resources. Assuming Huang's law, that's in 24 years. Assuming Moore's law, that's in 38 years.

With language models being the hot topic against vision models, nobody cares anymore about radiologists being potentially replaced. At this pace, we will be the last ones working on earth in a dark office with a computer and a big screen.

We are very pleased to announce Radiology: Artificial Intelligence's 2022 Impact Factor... 🌟 9.8 🌟 Sincere thanks to our authors, reviewers, editorial board, and RSNA Publications staff!

Facial de-identification in head CTs is an important step in #neurorad research. You can use the web app to de-identify your studies or run the code locally. Web app: huggingface.co/spaces/felipek… Code: github.com/kitamura-felip… #DICOM #ctscan #dataprivacy #healthcaretechnology #AI

GPT-4V Unseen territory in medical imaging. Radiology: Artificial Intelligence RSNA @CARadiologists Paper : arxiv.org/abs/2309.17421 (sections 4.1 p. 32-33, sections 9.3 p. 99-103)

Just received my cybertruck this morning.

I've read a ton of research papers this year. And to conclude this eventful year in AI, I've compiled a selection of 10 noteworthy papers from 2023 that I am discussing in my new article: magazine.sebastianraschka.com/p/10-ai-resear… To get a sneak peak, I'm covering: - insights into LLM training

Four international #radiology societies join RSNA to endorse a Special Report on #AI doi.org/10.1148/ryai.2…

Post-ImageNet, premature threats of AI replacing the apparently simple task of radiologists arose. Yet projecting 2D/3D imaging into nuanced language is far more complex than solving language alone with LLMs. Ultimately, radiology showcases the depth of human intelligence.

Comprehensive analysis of deep learning models on breast #cancer screening doi.org/10.1148/ryai.2… Alexandre Cadrin-Chênevert #breast #MachineLearning #mammogram