We show that learning models that optimize these metrics (instead of more standard ones such as MSE) leads to performance that aligns more closely with clinical needs. @AMIAinformatics@ShalitUri@DannyEytan
Excited to present our paper “Aiming for Relevance” at AMIA Informatics 2024! Our work introduces novel performance metrics for predicting ICU vital sign trajectories, inspired by clinicians’ empirical inputs about what matters most to them.
Breaking News: Just now, #Hamas snipers have reportedly killed dozens of children and women on the streets, targeting those attempting to travel from north to south #Gaza and those displaying white flags as a sign of peace. Similar acts have previously been attributed to Palestinian and #Hezbollah terrorists in Syria. They do not want civilians to leave; they want to use them as human shields and kill anyone who attempts to leave. Hamas terrorists in Gaza will, as usual, blame #Israel because it is easy and there is media that accepts this propaganda.
#HamasisISIS #HamasTerrorrists
Hamas in its own words:
October 7 is “just the first time and there will be a second, a third, a fourth…”
Hamas is openly admitting to the world that it will never abide by a ceasefire.
It will keep slaughtering Jews until the Jewish State itself is “annihilated.”
Why does the world expect Israel to sign a suicide pact and become the author of its own annihilation?
תראו: ככה מנצחים בויכוח. לא בצעקות. לא בנפנופי ידיים. לא בשאלות למראיינת מה היא הייתה עושה. אלא בציטוט מדויק, חוזר ונשנה, של עובדות, במקרה הזה של החוק הבינלאומי.
Surveillance camera footage from Netiv Haasara shows a large barrage of rockets being launched from northern Gaza, followed by a massive blast in the Strip, apparently caused by a failed projectile.
@AMIAinformatics I would appreciate it if you could clarify whether submitting a paper to the AMIA 2024 Informatics Summit’s paper track when some of its findings are planned to be included in a non-archival workshop with a different topic and emphasis is permitted.
In a nutshell, the already cognitively-taxed clinicians value predictions that are actionable on the short-term, and give new knowledge or insights: “tell me something I don’t know”.
I am happy to share that our paper “Tell me something interesting: Clinical utility of machine learning prediction models in the ICU” was accepted to @JBI_Journal! @ShalitUri@DannyEytan@ofraam
https://t.co/VysohdQw5W
The study also shows that not all errors are equal: The utility from correctly predicting vital signs depends on context: clinical norm, trend and surprise. (Figure : HR prediction utility vs. absolute values)