Greg Ver Steeg retweeted
✨🧠Excited to have had the opportunity to present our work titled 'Diffusion Bridge Models for 3D Medical Image Translation' in IEEE #embc2025 conference. Some of the important points of discussion from the paper are in this 🧵
Your diffusion model is secretly a noise classifier arxiv.org/abs/2407.08946
rdcu.be/dDm5k Nice work on latent space learning with biomedical data by amazing PhD student Myrl Marmarelis, who is working at the intersection of causality and machine learning. He's defending soon, so you should move fast if you want to lure him somewhere!
Greg Ver Steeg retweeted
1/8
🚀 AI Breakthrough: "Interpretable Diffusion via Information Decomposition" 🧠
- Quantitative understanding of conditional diffusion models.
- Align text-image data using mutual information.
- Goes beyond "attention".
🎉 Accepted at #ICLR2024!
Greg Ver Steeg retweeted
MI + diffusion! two new papers calculating (pointwise) mutual information as a difference of scores
cool experiment by @xxk0ng @gesteller identifying areas of an image relevant to changing a prompt
arxiv.org/abs/2310.07972
see also @michiard arxiv.org/abs/2310.09031
Greg Ver Steeg retweeted
Congrats to our graduating ISIers! #Classof2023
@harhrayr is earning his #PhD in #ComputerScience, which he obtained under the supervision of his two advisors @aram_galstyan & @gesteller.
Following graduation, Hrayr will join @Google Research in NYC as a #researchscientist!
Greg Ver Steeg retweeted
What are the #histopathology image #WSI patterns associated with gene expression in cancer? Can we discover them with #AI and use them for spatial profiling and #precision medicine?
Our preprint explores limitations & possibilities of #cpath for this.
biorxiv.org/content/10.1101/…
Robnik et al's Micro-canonical Langevin Monte Carlo arxiv.org/pdf/2303.18221.pdf sets a new standard for sampling. Elegant energy preserving dynamics that sample from a target distribution faster and with less bias!
Greg Ver Steeg retweeted
AM is a general approach for learning CNFs or SDEs, w/o backdrop thru dynamics in training
-natural for trajectory inference in biology, w/time snapshots given
-or generative modeling w/flexible endpoint dists or interpolating processes
(learned a ton from @k_neklyudov here :)
Happy to share an update to Action Matching arxiv.org/abs/2210.06662, a method that learns the evolution of any distributions
+ learning Fokker-Planck eq. for any evolution
+ trajectory inference for biology
+ quantum systems simulation
+ connections to OT (+ entropic, unbalanced)
Greg Ver Steeg retweeted
**AI can detect new materials**
New paper with Marcin Abram, @DingrevilleRemi, and @gesteller just published in NPJ Computational Materials rdcu.be/cV6q4
Slides borrowed from a GREAT talk @DingrevilleRemi will be presenting at a SIAM DS siammds22.us2.pathable.com/m…
🧵1/6
Greg Ver Steeg retweeted
Pattern formation is key in many physical and biological systems, but it can be hard to discern when transitions occur. We solve this challenge with self-supervised learning and neural nets to detect hierarchies of topological transitions. Details: rdcu.be/cV6q4
Greg Ver Steeg retweeted
Welcome to the official account of the Visual Intelligence and Multimedia Analytics Laboratory (#VIMAL) at @USC_ISI @USCViterbi @USC. Our research spans theoretical #AI, #ComputerVision, #misinformation #forensics, #deepfake detection and #biometrics.
Greg Ver Steeg retweeted
I'm happy to announce that I'm joining @Vanderbilt_CS this fall as an Assistant Prof, building a Machine Learning and Medical Imaging group (alongside some truly wonderful faculty already there).
Please reach out if you want to collab!
Greg Ver Steeg retweeted
We’ll be presenting our ICLR paper “Improving Mutual Information Estimation using Annealed and Energy-Based Bounds”, Tues @ 10:30am PDT/17:30 GMT
Talk: bit.ly/3vcFDW6
Paper: bit.ly/3Kcwi4X
w/@sicong_huang @MarzyehGhassemi @gesteller @RogerGrosse @AliMakhzani
Greg Ver Steeg retweeted
New Article: "A Metric Space for Point Process Excitations" by Marmarelis, Ver Steeg and Galstyan jair.org/index.php/jair/arti…
Greg Ver Steeg retweeted
Our manuscript is available here: arxiv.org/pdf/2108.13407.pdf
How do road networks evolve? I, with my colleagues Johannes Uhl, Stefan Leyk, and @KristinaLerman have explored this question by recreating road networks since 1900.
1/🧵
Accelerating graph learning apparenthorizons.com/2022/01…
