🔥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

Interesting deep learning work from my research mentors on detecting diabetes from smartphones, published in Nature Medicine today! Looking forward to future work in this field

Why has the United States handled this pandemic so badly? The Editors note that although we came into this crisis with enormous advantages, our current political leaders have demonstrated that they are dangerously incompetent.

New work using Radiopaedia.org to produce and test word embeddings, by Jae Ho Sohn, MD, MS et al from UCSF Imaging, published in J_Biomed_Info. Nice use of a publicly available corpus, and publication of over 1700 new radiology analogies for intrinsic validation. doi.org/10.1016/j.jbi.…

In our recent work in JAMA Cardiology we trained an #AI neural network for a wide variety of 12-lead ECG diagnoses that also “explains” its predictions using commonly available ECG data and labels. jamanetwork.com/journals/jamac… Read more: tison.ucsf.edu/explainable-ec…

The UCSF Center for Intelligent Imaging Big Data in Radiology (BDRAD) team have developed & evaluated an natural language processing (NLP) algorithm that matches clinical indications to appropriate AC guidelines. Find out how they did it ➡️ bit.ly/3rEMF3E

Great work by Gunvant Chaudhari, Timothy Chen, Yeshwant Chillakuru, Jae Ho Sohn, MD, MS, Christopher Hess, Valentina Pedoia, Thienkhai Vu & Youngho Seo on this approach that can be integrated into imaging ordering systems for automated access to guidelines. bit.ly/3rEMF3E

Highlighting some of the outstanding exhibits, posters and scientific talks by #UCSF Radiology residents, fellows and med students at #ASNR22. So fortunate to work with such talented and hard working physicians!

New blog post: 'A Natural Language Processing (NLP) Search Algorithm to Facilitate Use of Appropriateness Criteria' bit.ly/3Gjwr6x via UCSF Center for Intelligent Imaging

Speech recognition errors during radiology report dictation are often not caught by traditional spell checkers. In Radiology: Artificial Intelligence, we evaluate a context based deep learning-based approach using #BERT to detect dictation errors and suggest corrections.

"Our algorithm provides more relevant results than a custom Google search engine, especially for longer and more clinically complex imaging indications," says Gunvant Chaudhari, UCSF School of Medicine student & first author on this recent UCSF Center for Intelligent Imaging BDRAD study. bit.ly/3Gjwr6x

A deep learning model evaluating myelination patterns predicted gestationally corrected age of neonates and infants based on brain MRI scans. Joshua Chen Gunvant Chaudhari Christopher Hess Andreas Rauschecker Yi Li UCSF Imaging UCSF Center for Intelligent Imaging bit.ly/3OiJJlZ

A #DeepLearning model that evaluates myelination patterns can predict the gestationally corrected age of neonates & infants on the basis of T1- and T2-weighted MRI scans of the brain. bit.ly/3be77n6 Great work Andreas Rauschecker Yi Li & team!

Note: my session has been rescheduled to Tues, 11/29, from 3:00-4:00pm. But there’s another great #NLP session during the same time block on Mon, 11/28, led by Jae Ho Sohn, MD, MS Gunvant Chaudhari Timothy Chen UCSF Center for Intelligent Imaging UCSF Imaging #RSNA22

Lucky to be working with these talents on NLP deep learning lab and similar image retrieval projects. More to come! UCSF Imaging UCSF Center for Intelligent Imaging UCSF Cardiothoracic Imaging #UCSFproud

I can't recall a UC San Francisco grand rounds that was more fascinating or mind-blowing than this far-ranging discussion of #ChatGPT's implications for healthcare – clinical, research, and education – by Atul Butte, Aaron Neinstein, MD, Sara Murray & Dan Lowenstein. youtube.com/watch?v=j-aOCu…

ChatLLaMA - an open-source implementation of LLaMA based on RLHF. Claims a 15x faster training process than ChatGPT. It allows users to fine-tune personalized ChatLLaMA assistants. github.com/nebuly-ai/nebu…

Geoff Tison UCSF Cardiology created an AI-powered algorithm that accurately predicts 💝myocardial injury using #ECG data. ✅The algorithm's negative predictive value of 80% allows it to triage low-risk patients using a simple and cost-effective ECG. rdcu.be/c6tCu #AI

UCSF medical student and 0.55T lung MRI researcher, Felicia Tang, commenting on her recent research and potential future directions for lung MR: auntminnie.com/resources/conf…

T1. BERT is a type of large language model that takes into account context of words in a sentence when doing natural language processing. It was a significant leap and initial step breakthrough towards the current large language models that we see in the field. #RadAIChat

Can smarter search boost diagnostic AI? 🧠 A study by Cody Savage, Gunvant Chaudhari, Andrew Smith & Jae Ho Sohn, MD, MS shows that RadSearch, a specialized semantic search model, improves radiology report retrieval and boosts LLM diagnostic accuracy. See how it works! pubs.rsna.org/doi/10.1148/ra…