Scott McGrath
banner
smcgrath.phd
Scott McGrath
@smcgrath.phd
Biomedical Informatics PhD • CITRIS Health @UC Berkeley • FAMIA • Focusing on Informatics and AI in medicine • Linfield U. Grad • Missoula MT

https://citris-uc.org/people/person/scott-mcgrath/
Pinned
Hello #AMIA2025, so glad to see all 2,145 of you here in Atlanta .

I wanted to give you a guide on how to make the most of the conference here on Bluesky, and hopefully convince you to continue the conversation after you head home.

Here are some tips I want to share in a thread 🧵 1/10
A Guide to Bluesky for the AMIA Community
Welcome to Bluesky!
amia.org
Reposted by Scott McGrath
Is "Agent Engineering" the next big thing? Getting so-called "AI Agents" to work together effectively isn't trivial. @moorejh.bsky.social and I discuss the need for rigorous development in our latest editorial in BioData Mining

biodatamining.biomedcentral.com/articles/10....
From prompt engineering to agent engineering: expanding the AI toolbox with autonomous agentic AI collaborators for biomedical discovery - BioData Mining
biodatamining.biomedcentral.com
November 16, 2025 at 5:15 PM
Hello #AMIA2025, so glad to see all 2,145 of you here in Atlanta .

I wanted to give you a guide on how to make the most of the conference here on Bluesky, and hopefully convince you to continue the conversation after you head home.

Here are some tips I want to share in a thread 🧵 1/10
A Guide to Bluesky for the AMIA Community
Welcome to Bluesky!
amia.org
November 16, 2025 at 8:09 PM
#AMIA2025 Symposium has 2,145 people here!

Of 1,977 submissions, overall acceptance rates included:
Panels at 37%,
Regular Papers at 43%,
and Posters at 61%.

A strong presence of student and physician coauthors was called out.
November 16, 2025 at 6:59 PM
Reposted by Scott McGrath
At the #AMIA2025 orientation. Come and meet us and learn more about @amianiwg.bsky.social !
November 16, 2025 at 5:48 PM
Congratulations @billhersh.bsky.social

Well deserved!

#AMIA2025
November 16, 2025 at 6:46 PM
November 16, 2025 at 6:45 PM
Really good discussion in Leveraging Artificial Intelligence for Evidence Synthesis: Screening, Extraction, and Synthesis workshop.

Just sharing one of the tools I mentioned. Asta, which a really interesting AI powered academic paper finder.

asta.allen.ai/chat

#AMIA2025
Ai2 Asta
A scholarly research assistant with broad and deep coverage via a corpus of 108M+ scholarly abstracts and 12M+ full text papers. A project from Ai2.
asta.allen.ai
November 16, 2025 at 3:59 PM
For the next step of the workshop: The goal of data synthesis is to transform extracted data into meaningful insights. AI can accelerate this by suggesting themes, but humans provide the interpretive rigor: verifying relevance, interpreting meaning, and ensuring transparency.
#AMIA2025 #AcademicSky
November 16, 2025 at 3:36 PM
Continuing the workshop: AI can assist data extraction by organizing info into structured tables and batch-extracting fields from multiple sources.
Human reviewers remain essential. They must verify the extracted data for accuracy, interpret context, & resolve discrepancies.
#AMIA2025 #AcademicSky
November 16, 2025 at 2:55 PM
Reposted by Scott McGrath
#AMIA2025 Come meet Abby Kietzman @abbykietz.bsky.social and learn about the Reasearch Informatics Navigator role at Children's Mercy @research.childrensmercy.org . Poster #P78 Sunday Nov. 16.
November 16, 2025 at 2:19 PM
Reposted by Scott McGrath
Day 2 of the #AMIA2025 Symposium! On deck: AI-enabled health IT applications can cause considerable cascading effects in the clinical processes, workflow, and consequently, on throughput and efficiency. How can we leverage these tools to help us take better care of patients? #MedSky #patientsafety
November 16, 2025 at 2:30 PM
Evidence synthesis is essential, but repetitive work. AI can assist by generating preliminary classifications and summarizing abstracts, but human reviewers must verify recommendations, interpret context, and resolve ambiguities. #AMIA2025 #AcademicSky
November 16, 2025 at 1:52 PM
Next up in Leveraging AI for Evidence Synthesis: Screening, Extraction, and Synthesis workshop by Christie Martin, Grace Gao, Jenna Marquard, @scottsittig.bsky.social, Vishala Mishra, Julie Doberne, Zainab Balogun, Sayantani Sarkar, Ming-tse Tsai, Akshitha Gopikrishnan, & Caitlin Bakker #AMIA2025
November 16, 2025 at 1:24 PM
Super fun! Enjoyed making a new friend and thanks to @drdannysands.bsky.social for leading it as always! #AMIA2025
👟 What a way to start the day at #AMIA2025! This morning's Fun Run/Walk brought together informaticians for 2-4 miles of fresh air, exercise, and great conversations. There's something special about connecting with colleagues in such an energizing environment.

Thanks to everyone who joined us!
November 16, 2025 at 12:30 PM
Reposted by Scott McGrath
Thrilled to see our primary and secondary faculty sharing their expertise at #AMIA2025! From AI in health to genomics and clinical informatics, their work is shaping the future of healthcare.

Check out their sessions and join the conversation ➡️ https://cudbmi.info/amia-2025-1
November 15, 2025 at 7:55 PM
Part of the background info for SL and RL of LLMs for clinical applications: General LLMs lack medical grounding. SFT adapts them for healthcare using instruction-response pairs. This creates a specialized model and reduces errors.
#AMIA2025 #MedSky #MLSky
November 15, 2025 at 6:16 PM
Starting in in the next workshop: A Hands-on Tutorial on Supervised Fine-Tuning and Reinforcement Learning of Large Language Models for Clinical Applications
#AMIA2025 #MLSky
November 15, 2025 at 6:03 PM
Reposted by Scott McGrath
This year marks my 40th consecutive #AMIA2025 fall meeting, dating back to SCAMC, and including the virtual meeting during the pandemic in 2020. Extra special this year is my daughter Alyssa attending, along with my grandson, and we will all be decked out in AMIA t-shirts!
November 15, 2025 at 3:51 PM
Reposted by Scott McGrath
AGI = avoiding grammar issues
AGI any day now
November 15, 2025 at 11:34 AM
Dr. Adela Grando provides an introduction to foundational ethical AI frameworks. Really valuable element of this workshop. If you want to deploy AI ethically, these are guides on how you can do that.
#AMIA2025 #MLSky #MedSky 1/3
November 15, 2025 at 3:08 PM
Reposted by Scott McGrath
1st day at the @amia.org Annual Symposium here in ATL off to a great start! Attending the Workshop on Interactive Systems in Healthcare! Let's get to flourishing! #WISH #healthAI #wellbeing
November 15, 2025 at 2:17 PM
Dr. Raju Gottumukkala discusses poor AI consent and privacy design treats people as data sources to be extracted. A diabetes app user may agree to a 40-page document, unaware their data is shared and retained indefinitely. This exploits vulnerable users, as consent is not truly voluntary. #AMIA2025
November 15, 2025 at 2:40 PM
Dr. Tsai leads a talk about the rationality and irrationality of human decisions, starting with the average human makes 35k decisions per day (both rational and irrational). What happens to our decisions when AI is introduced. #AMIA2025 #MLSky
November 15, 2025 at 2:33 PM
When discussing the topic of representativeness of AI training datasets, Dr. Jenna Marquard asks those in attendance, how many people are using a health devices with AI functionality, and then asked if anyone knows what training set that AI was built on, or where to find it. #AMIA2025
November 15, 2025 at 2:20 PM