💥 Happening in 10 days! 📷 Francesco Di Giovanni and Kacper Kapuśniak will talk about their recent paper on Metric Flow Matching (MFM), a novel generative approach using data-induced Riemannian metrics for more accurate interpolations. Hosted by Michael Bronstein and Bruno Correia 📅 Fri, Jul

Reading group session tomorrow will be about this gem! "Metric Flow Matching for Smooth Interpolations on the Data Manifold" arxiv.org/pdf/2405.14780 I am sure these data dependent metrics they use could be useful elsewhere Join on zoom Monday at 11am ET: portal.valencelabs.com/logg

Francesco Di Giovanni and Kacper Kapuśniak's talk on Metric Flow Matching is now live on our YouTube channel and here on X

Here is our metric flow matching discussion video with Francesco Di Giovanni and Kacper Kapuśniak. youtu.be/6cbOim_HjPg?fe…

🦕Proving Approximation Results with minimal assumptions in VP-SDE Diffusion. Explore how SDE choices impact sample quality and gain insights into Score Matching in Denoising Diffusion Models! 🦕 Link: arxiv.org/pdf/2305.09605… Presented at UAI2024!

I am happy to announce that our paper called Metric Flow Match (MFM) has been accepted to NeurIPS Conference #neurips2024 🔥🔥🔥 I am grateful to all my collaborators and see u in Vancouver! ⛰️ Teo Reu Alex Tong Michael Bronstein Joey Bose Francesco Di Giovanni

Got very lucky today, 4/4 papers accepted to #NeurIPS2024 NeurIPS Conference !!!! 3 posters and 1 spotlight! Extremely grateful to my collaborators, especially the junior authors who really pulled this off! Kacper Kapuśniak Oscar Davis Teo Reu Damien Ferbach and of course my awesome

I am top 2% researcher according to my mom.

We are happy to announce the Oxford LoG meet-up, LoG-Ox! The event is free to attend and will happen on November 25th. We plan to have keynote talks, lightning talks, posters and socials. More information and signup form is available here: log-ox.github.io. Learning on Graphs Conference 2024

Camera-ready version and code available for MFM! 📷 📜arxiv.org/abs/2405.14780 💻github.com/kksniak/metric…

Christian S. Perone I think you'd enjoy our Metric Flow Matching Paper that makes learning those data-dependent geodesics a bit faster 😃 arxiv.org/pdf/2405.14780

Valence Labs will be co-hosting a TechBio social with Recursion and NVIDIA at #NeurIPS in Vancouver. Join us on Thurs, Dec 12th. RSVP here: lu.ma/biikt7ox Our team will also be at NeurIPS throughout the week. See below for a summary of our papers👇

Oscar Davis Metric Flow Matching For Smooth Interpolations on the Data Manifold, led by Kacper Kapuśniak arxiv.org/abs/2405.14780

Metric Flow Matching for Smooth Interpolations on the Data Manifold Kacper Kapuśniak Peter Potaptchik Teo Reu Leo Zhang Alex Tong Joey Bose Francesco Di Giovanni nips.cc/virtual/2024/p… x.com/KKapusniak1/st…

At #NeurIPS and interested in Flow Matching? Come and find us on Thursday presenting MFM - a generalization of FM with probability paths supported on the data manifold 🗺️ Poster: #2411 East Exhibit Hall A-C ⏲️Thu 12 Dec 11 am-2pm neurips.cc/virtual/2024/p… Reach out to chat!

We're presenting MFM right now! Come check it out in East Ballroom poster # 2411 🩷🩷🩷 Teo Reu Joey Bose Alex Tong Michael Bronstein Francesco Di Giovanni

🍩 Topological blindspots is coming to ICLR as an oral presentation! 🍩 We prove that message-passing based topological deep learning (TDL) architectures are unable capture basic topological invariants including homology, orientability, planarity and more.

Excited to share that new work on Curly Flow Matching with my amazing collaborators Lazar Atanackovic Kacper Kapuśniak Michael Bronstein Joey Bose Alex Tong has been accepted to Learning Meaningful Representations of Life and MLGenX Workshop @ ICLR 2024 ICLR 2026 workshops!🧬✨

🚨 ICML 2025 Paper 🚨 "On Measuring Long-Range Interactions in Graph Neural Networks" We formalize the long-range problem in GNNs: 💡Derive a principled range measure 🔧 Tools to assess models & benchmarks 🔬Critically assess LRGB 🧵 Thread below 👇 #ICML2025

Flow Matching models often struggle to balance memorization and generalization. 😱 We set out to fix this — by using the geometry of the data manifold. Introducing Carré du Champ Flow Matching (CDCFM)🧑‍🎨🥖 — improving generalization without sacrificing sample quality.