🎉 Update: This work got accepted to #icml2025!!. Huge thanks to my amazing co-authors @LorenzLinhardt, Marco Morik, @jdppel, @skornblith, and @lukas_mut for their great work and to all collaborators! 🙏. 📄 Paper: 💻 Code: 🧵1/3

The core message remains: Training objective drives the consistency of pairwise representational similarities, and similarity-performance correlations depend on the dataset structure. 🧵3/3.

Join millions who have switched to Grok.

During the rebuttal, we've added: improved local structure measurements, a CKA stability analysis, and additional dataset observations. 🧵2/3.

RT @fabian_theis: 1/🚀 Excited to share RegVelo, our new computational model combining RNA velocity with gene regulatory network (GRN) dynam….

RT @val_boeva: 🚀 New preprint from our lab, @ekrym2 and @fabian_theis : UniversalEPI, an attention-based method to predict enhancer-promote….

RT @fabian_theis: 1/ 🧬 Single-cell genomics reveals biological variations beyond cell types. Unveiling these in separate latent dimensions….

RT @hense96: ❓ Do histopathological foundation models eliminate batch effects? ❓. The surprisingly clear answer is: they do not!. Find out….

Want to learn more? Read our new preprint:✨.Deeply grateful to @LorenzLinhardt, Marco Morik, @jdppel, @skornblith, and @lukas_mut for their rigorous work and dedication throughout this project. Each contribution was essential to make this possible! 🙏 7/7.

2nd key insight: The link between model similarity & behavior varies by dataset. Single-domain sets show strong correlations, while some multi-domain ones have high-performing, dissimilar models. Thus, the Platonic Representation Hypothesis may depend on the dataset's nature. 6/7

Key finding: Training objective is the crucial factor for similarity consistency! SSL models show remarkably consistent representations across stimulus sets compared to image-text and supervised models, which show high variance in their consistency due to dataset dependence. 5/7

Thus, we suggest a framework to systematically study if relative representational similarities between models remain consistent. We measure similarities between sets of models with different traits and their correlation across dataset pairs to assess stability across stimuli. 4/7

First finding: Representational similarities do not transfer directly across datasets, showing high variability across datasets, such as different ranges and patterns. 3/7

The Platonic Rep. Hypothesis by @phillip_isola et al. suggests foundation models converge to a shared representation space. Yet, most studies consider single datasets when measuring representational similarity. Thus, we were wondering: Does this convergence hold more broadly? 2/7.

If two models are more similar to each other than a third on ImageNet, will this hold for medical/ satellite images? Our preprint analyzes how vision model similarities generalize across datasets, the factors that influence them, and their link to downstream task behavior. 🧵1/7.

RT @JoYatesResearch: New Preprint 🚨: Our lab finds that high mitochondrial RNA (pctMT) levels, typically flagged as cell death in single-ce….

RT @hense96: Big thanks to my collaborators and mentors, first and foremost @IdajiMina, as well as @EberleOliver, @thomschnake, @jdppel, @l….

RT @lukas_mut: 🚨thingsvision ( has a few new gems for you:. - 🔥efficient mini-batch feature extraction for using y….

RT @stevain: First publication of my postdoc in Berlin is out on arXiv today! 🎉. ~~> We use diffusion probabilistic models to relax distort….

This work was made possible by the amazing team with @KraftAgnieszka, @flobarkmann, @JoYatesResearch, and supervisor @val_boeva. Thank you all!.

Elevate single-cell scoring accuracy using ANS, a reliable method compatible with Scanpy & Seurat. We even reimplemented UCell & Jasmine in Python! 🚀Check out our preprint: .GitHub: #SingleCellAnalysis #CancerResearch #DataScience.