As the Iranian regime imposes a forced forgetting of the massacres in January, it has begun targeting not only wounded protesters but medical workers, who have borne witness to some of the worst atrocities. https://t.co/ZYA12KWwdn
I wonder if if anyone inside Iran sees my posts, if they can connect at all.
If you are there and see these, I hope they give you a little encouragement.
At least just to know that you are not alone.
Free Iran!
1 in 3 adults have varicose veins, putting them at risk of chronic venous insufficiency. A new interpretable AI model focuses risk assessment of endovenous thermal ablation, one of the main treatments for CVI that can cause complications like blood clots: https://t.co/tWuN2vdZf4
🧠What if your risk score came with a why?
Our interpretable model can predict if and why a patient is at risk of DVT after undergoing EVTA.
Personalized, explainable, actionable.
Read more DOI: 10.1016/j.jvsv.2025.102253
@yuma_or@hayfarani@dbertsim@DaniaDaye@AIHealthMIT
Coming a full circle 10 years later - I was mentored by @PennRadiology #MitchSchnall for my first @RSNA medical student grant in 2013 and today I am mentoring @Azitabari on her first #RSNA23 grant.
Thank you @RSNA for the prestigious Seed Grant award and my first funding as a PI!
And huge thanks to my forward-thinking mentors @DaniaDaye and @Mike_Gee8 for always leading with integrity, empowerment and wisdom.
We build a lie detector for large language models and show that they represent uncontroversially true and false sentences on opposite sides in a linear space. This is IMHO further evidence they they’re not just overhyped “stochastic parrots”: