Twitxy
Saved
@AdrianDalca's profile picture
View on X

Adrian Dalca

@AdrianDalca

Assistant Professor, MGH, Harvard Medical School;
Research Scientist, MIT

AI/Machine Learning, with focus on Health and Vision
Cambridge, MA adalca.mit.edu Joined September 2014
182 Following 1.4K Followers 816 Posts

Latest posts 5

Adrian Dalca @AdrianDalca
@MarianneRakic is presenting on a cool new idea -- probabilistic in-context segmentation -- at ICCV workshops.

Make sure to catch up with her during the conference as well!
Marianne Rakic @MarianneRakic Oct 1, 2023
Very excited to be at #ICCV23 in Paris Mon-Fri 🥳
Are you interested in uncertainty quantification for generative models, in-context learning for vision, or multi-annotator data too?
I would love to chat 🤩

I am presenting a poster at #UnCV and at #WiCV 🙌
View quoted post on X
Adrian Dalca @AdrianDalca
@VictorButoi and @jjgort are at ICCV -- go to their UniverSeg poster on Friday!

And contact them directly or stop them in the hallway to chat with them :)
Victor Butoi @ion_barrel Sep 29, 2023
Very excited to be Paris for my first conference in #ICCV23 🎉.

Please let me know if you'd like to chat 😄! (ESPECIALLY if you're also excited about in-context learning for vision, statistical inference for DNNs, uncertainty, or understanding generative image models.)
View quoted post on X
Adrian Dalca @AdrianDalca
UniverSeg can solve new biomedical segmentation tasks at inference, without any re-training!

Leads @VictorButoi & @jjgort will present at #ICCV2023!

Code, tutorial, wts @ universeg.csail.mit.edu

w/ T. Ma, @mertrory, J. Guttag. @MIT_CSAIL, @MIT, @MGHMartinos, @FreeSurferMRI
Victor Butoi @ion_barrel Aug 29, 2023
Presenting UniverSeg: an in-context learning model for medical image segmentation to appear at #ICCV2023 🎉! (w/ @jjgort, @mertrory, @AdrianDalca, and others)

@MIT_CSAIL, @MIT, @MGHMartinos

🧵1/N (project-page, demo, and paper links 🔗at the end)
View quoted post on X
1 replies 2 reposts 29 likes 3.7K views View on X
Adrian Dalca @AdrianDalca
The vitriol really is unacceptable, especially from senior people. Juvenile, really.

Whatever the critique (and I do think it's a problem and the paper should not have been uploaded) there's ways to do it that's productive, kind, and didactic (like @sauhaarda et al response)
Victor Butoi @ion_barrel Jun 17, 2023
The very extreme reactions to the preprint of GPT4 ability on MIT curricula is frankly appalling.

Don't demonize undergrads who are just learning how to do research and also... it's just not a big deal? This isn't a publication, and the authors will improve it.
View quoted post on X