erkin ötleş
@eotles.bsky.social
MD-PhD. emergency resident doc/engineer. designing / evaluating / implementing #AIinMedicine. hopefully bridging the gap between physicians & engineers 🤞🏾
🚨 new JAMIA Open paper on how we define a “fall” in the ED using EHR data
our team ran an NLP algorithm over 50k+ ED provider notes and compared results to diagnosis code-based definitions
nearly half of the falls found by NLP weren’t captured by codes at all 🧵
our team ran an NLP algorithm over 50k+ ED provider notes and compared results to diagnosis code-based definitions
nearly half of the falls found by NLP weren’t captured by codes at all 🧵
Evaluation of falls detected by natural language processing algorithm and not coded external cause of morbidity
AbstractObjective. Falls are a leading cause of morbidity and mortality among older adults. Common methods for identifying fall-related ED visits within bo
academic.oup.com
June 23, 2025 at 4:59 PM
🚨 new JAMIA Open paper on how we define a “fall” in the ED using EHR data
our team ran an NLP algorithm over 50k+ ED provider notes and compared results to diagnosis code-based definitions
nearly half of the falls found by NLP weren’t captured by codes at all 🧵
our team ran an NLP algorithm over 50k+ ED provider notes and compared results to diagnosis code-based definitions
nearly half of the falls found by NLP weren’t captured by codes at all 🧵
🚨💻🦠💩
our AI-guided system for C. difficile infection (CDI)
prevention study is now live in
@jamanetworkopen.com
new research by: @umichmedicalschool.bsky.social
jamanetwork.com/journals/jam...
🧵 key takeaways from developing + evaluating this real-world hospital AI implementation:
our AI-guided system for C. difficile infection (CDI)
prevention study is now live in
@jamanetworkopen.com
new research by: @umichmedicalschool.bsky.social
jamanetwork.com/journals/jam...
🧵 key takeaways from developing + evaluating this real-world hospital AI implementation:
Clostridioides difficile Infection Prevention in a Hospital Setting With AI
This quality improvement study examines a 28-month quality improvement program that used an institution-specific AI model to guide Clostridioides difficile infection prevention efforts at a large acad...
jamanetwork.com
June 18, 2025 at 7:39 PM
🚨💻🦠💩
our AI-guided system for C. difficile infection (CDI)
prevention study is now live in
@jamanetworkopen.com
new research by: @umichmedicalschool.bsky.social
jamanetwork.com/journals/jam...
🧵 key takeaways from developing + evaluating this real-world hospital AI implementation:
our AI-guided system for C. difficile infection (CDI)
prevention study is now live in
@jamanetworkopen.com
new research by: @umichmedicalschool.bsky.social
jamanetwork.com/journals/jam...
🧵 key takeaways from developing + evaluating this real-world hospital AI implementation:
@doximity.bsky.social what’s up with your AI scribe? the promo materials make me want to try it out!
April 4, 2025 at 1:12 PM
@doximity.bsky.social what’s up with your AI scribe? the promo materials make me want to try it out!
you mean so much to me #ai #valentines
February 14, 2025 at 12:11 PM
you mean so much to me #ai #valentines
kiddo: i can count to 0
me: go for it
kiddo: oneee
me: 🤪
me: go for it
kiddo: oneee
me: 🤪
February 4, 2025 at 12:17 AM
kiddo: i can count to 0
me: go for it
kiddo: oneee
me: 🤪
me: go for it
kiddo: oneee
me: 🤪
Reposted by erkin ötleş
LLMs, despite immense potential, fail at basic reasoning tasks like loading the dishwasher.
January 17, 2025 at 5:15 AM
LLMs, despite immense potential, fail at basic reasoning tasks like loading the dishwasher.
it doesn’t make sense to have a single threshold for AGI. human intelligence is wide, varied, & unassured. why should AI be any different?
instead of an fixating on arbitrary definition, we could focus on understanding & improving performance across spectra of useful tasks
instead of an fixating on arbitrary definition, we could focus on understanding & improving performance across spectra of useful tasks
January 14, 2025 at 6:26 PM
it doesn’t make sense to have a single threshold for AGI. human intelligence is wide, varied, & unassured. why should AI be any different?
instead of an fixating on arbitrary definition, we could focus on understanding & improving performance across spectra of useful tasks
instead of an fixating on arbitrary definition, we could focus on understanding & improving performance across spectra of useful tasks
“one often hears a commentator call out, in hope, for the arrival of someone to serve as ‘the Newton of the mind.’ … the hoped-for Newton has already come and, alas, already gone. His name is John von Neumann”
December 31, 2024 at 12:45 AM
“one often hears a commentator call out, in hope, for the arrival of someone to serve as ‘the Newton of the mind.’ … the hoped-for Newton has already come and, alas, already gone. His name is John von Neumann”
is the only difference between “gestalt” and “unexplainable machine learning” the involvement of humans?
December 17, 2024 at 6:04 PM
is the only difference between “gestalt” and “unexplainable machine learning” the involvement of humans?
