Mahdi Soltanolkotabi
@mahdisoltanol
Followers
563
Following
1K
Statuses
102
work on foundations of AI, MLLM reliability/Eval, optimization, probability/stats, AI 4 science/healthcare; Prof & director of center on AIF4S @USC 🚲🏔️🥾🏊♂️
San Francisco, CA
Joined August 2022
🚨 Introducing MediConfusion: A new challenging VQA benchmark for Medical MLLMs! 🚨 All available models score below random guessing on MediConfusion, raising serious concerns about their reliability for healthcare deployment. with @shahabsepehri @zalan_fabian @MSKMarSol 🧵 1/6
1
9
42
RT @AndrewHires: Great, detailed explainer. Indirect costs ARE research costs, but are ones that are hard to assign to individual projects.…
0
4
0
RT @robinomial: Our work (with @tianyi_zhou12 @DeqingFu and Vatsal Sharan) published at NeurIPS 2024 already found that pretrained LMs do a…
0
9
0
RT @tianyi_zhou12: Great to see others discovering similar findings as we did in our Neurips2024 paper (. We call t…
0
4
0
@AndrewHires @LarsLindemann2 Yes 😅 full disclosure: we don’t know much about crypto. Our image recon architecture basically seems to work here as well …
0
0
0
@AndrewHires @LarsLindemann2 Hope u can go back soon and Palisades gets rebuilt quickly 🙏 ❤️. In the mean time Mandy repeats!
0
0
0
@docmilanfar 😅 full disclosure ATM one of said folks (in SLC feeling the full force of the polar vortex)
0
0
2
Great to have one of our own @USCMingHsiehEE grads* @DimitrisPapail back for an awesome talk on teaching algorithms via self improvement. More detail
1
0
31
RT @HaoLi81: For those who are from the malibu/pacific palisades/santa monica/brentwood area, here is a 400 Megapixel photo of the area. In…
0
4
0
Nice blog and talk by @beenwrekt at @SimonsInstitute. I did try Coach Ben’s Program but reached the exhaustion phase by day II.
Getting a little speculative today: How some necessary properties of homeostatic stress response might connect to progressive overload.
0
2
14
RT @max_simchowitz: A very exciting personal update: In January, I’ll be joining @CMUMLD as tenure-track assistant professor! My lab will f…
0
42
0
RT @HannaHajishirzi: Thrilled to introduce Tülu 3: a family of open state-of-the-art post-trained models with full access to data, training…
0
15
0