Twitxy Saved

Adarsh Subbaswamy

@_asubbaswamy

Using AI to improve and accelerate medical product development | Assistant Professor @UMBaltimore | CS PhD from @JohnsHopkins
asubbaswamy.github.io Joined August 2017
256 Following 424 Followers 171 Posts

Latest posts 5

Adarsh Subbaswamy @_asubbaswamy
It turns out monitoring the performance of ML models is hard, especially in settings like healthcare where model predictions affect user decisions

If you're at NeurIPS check out Jean's talk and we'd love to hear your thoughts re monitoring
Jean Feng @Jean_J_Feng Dec 16, 2023
Our framework brings together ideas from causal inference and statistical process control. Please drop by if you're interested! Preprint here: arxiv.org/abs/2311.11463 It was wonderful working on this with coauthors from UCSF and FDA! @pirracchior @_asubbaswamy @harvineet_singh
View quoted post on X
1 replies 0 reposts 12 likes 1.7K views View on X
Adarsh Subbaswamy @_asubbaswamy
Check out the beginning of a series of posts on external validity. Excited to read the next ones.

See also some response threads from coauthors and me:
x.com/_asubbaswamy/s…
x.com/kdpsinghlab/st…
x.com/IAmSamFin/stat…
Excited to publish the next part of me & @beenwrekt's blog series!

Often, we inappropriately characterize *any* failure of deployed ML models as a "distribution shift". This both muddles our understanding & limits our vocab of the actual issues we see.

argmin.net/2022/03/31/ext…
View quoted post on X
Adarsh Subbaswamy @_asubbaswamy
Overall agree with the thread's point that grad students need to be willing to repeatedly take feedback and improve their work. But I think the focus on "work is always unfinished" really misses the mark, especially in the context of higher education. 1/7
César A. Hidalgo Verified account @cesifoti Sep 12, 2021
At that point, I find it useful to have this conversation:
“Look, the paper will need to go through 20 to 30 iterations to be completed. If each iteration takes 2 months, then we will be ready in more than three years. It is up to you.” 7/N
View quoted post on X
Adarsh Subbaswamy @_asubbaswamy
In this NEJM letter we consider how dataset shift interacts with clinicians: What does it look like in practice? How can we identify and mitigate it?

Collaboration with a stellar, interdisciplinary team.
Sam Finlayson Verified account @IAmSamFin Jul 14, 2021
Letter: "The Clinician and Dataset Shift in Artificial Intelligence" out in @NEJM w/ @_asubbaswamy , @suchisaria, @kdpsinghlab, @zittrain, @zakkohane, et al

Link: nejm.org/doi/full/10.10…

Full version of Table (w/ details+citations): sgfin.github.io/assets/dataset…

Brief thread: 1/n
View quoted post on X