Mathilde Papillon🦋 mathildepapillon .bsky .social

TopoTune is going to ICML 2025!🎉🇨🇦 Curious to try topological deep learning with your custom GNN or your specific dataset? We built this for you! Find out how to get started at geometric-intelligence.github.io/topotune/

@ninamiolane @gbg1441 @HaewonJeong00 @geometric_intel @ai_ucsb 🚀 Build fairer personalized ML! Our work introduces tools to evaluate trade-offs in accuracy, fairness & explanation quality. 📷Read “When ML Gets Personal” here: [arxiv.org/abs/2502.02786]

Super excited to announce that our paper, “When Machine Learning Gets Personal”, has been accepted to #ICLR2026! 🇧🇷 We propose a unified framework to reliably quantify how personalizing a model influences both prediction accuracy and explanation quality. arxiv.org/abs/2502.02786

Come join us at SOhO restaurant tonight. I will give a public talk about the newly emerging field of "physics of intelligence" Abstract below!

🏆UCSB ranked #1 public university in the US for scientific impact, and 3rd in universities overall!🎓 Congratulations to all of our researchers who are leading the global conversation.🌐🧵 @ucsantabarbara @UCSBengineering @UofCalifornia

As students tap into #AI tools that can teach better than their instructors, what does the future of education look like? 👉 @AbbyBertics from our lab @geometric_intel explores this question for the @TheEconomist !

for the first time, eeml treads mountainous terrain 🏔️ while being very close to the seaside 🌊 not to mention it's the best AI summer school in EE & beyond! come join us, it'll be amazing! @clarelyle @dorothychou @ninamiolane @sedielem @apsarathchandar @re_rayne @fedzbar

Of course, this wouldn't have been possible without the incredible support and hard work of the core team. Huge thanks to @gbg1441, @mathildepapillo, and @ninamiolane for driving TopoBench forward this year!

2025 has been a landmark year for TopoBench, with the great achievement of surpassing 200 stars this month! github.com/geometric-inte… Time to recharge for the new year. We will be back soon with even more cutting-edge research and ambitious initiatives. Happy holidays!

Thrilled to see our team pushing the boundaries of AI in neurodegeneration research. Amazing work using GNNs + LLMs to uncover protein drivers of cognitive decline in FTD, read more here👇

What a great time at the @neur_reps workshop this year! We had an INCREDIBLE community of researchers from AI, neuroscience, physics, and mathematics come together to discuss their work on mathematical structure of representations in artificial and natural neural systems 🧠🤖 Thanks to all authors for submitting record number of abstracts and proceedings papers! The accepted papers will be available in PMLR soon. Thanks to our amazing speakers and panelists @tegmark, @s_y_chung, @SuryaGanguli, @kfrankelab, @klindt_david, @YueSong48287250, @t_andy_keller, and Razvan Pascanu for joining us! and thanks to our sponsors @newtheoryai for a great after party 🥳 Honored to work with a great team @simoneazeglio, @cashewmake2, Chase van de Geijn, @naturecomputes, and @ninamiolane!

Using explainable GNNs on proteomics, we predict cognitive decline in FTD and use LLMs to map key proteins onto the scientific literature. Excited to see this work in @TheTransmitter thetransmitter.org/machine-learni… @ninamiolane @RowanSaloner @kbcasaletto @gbg1441

"The Algebra of Spatial Navigation" Joint work w/ @KuninDaniel Giovanni Marchetti, @hopfbifurcator @AdeleMyersPhD Chris Kymn, Feng Chen, Dhruva Karkada, James Simon, Michael DeWeese, @SuryaGanguli "Topological Deep Learning" Joint work w/ @mathildepapillo @gbg1441 @lev_telyatnikov @naturecomputes @ClaBat9 Florian Franzen, Martin Carrasco, @mustafahaji Marco Montagna, Michael Schaub, @theopapamarkou @tolga_birdal & many others @PyT_Team_

Check out our latest work on Geometric Machine Learning at #NeurIPS this week. We are recruiting PhD students and postdocs this cycle — please reach out if you are interested in joining us. Main conference: Higher-Order Learning with Graph Neural Networks via Hypergraph Encodings led by Raphael Pellegrin, Lukas Fesser Fri, 12/5, 4:30 PM – 7:30 PM, Exhibit Hall C,D,E #3804 openreview.net/pdf?id=oeMK0Js… Workshops: Towards distillation guarantees under algorithmic alignment @diffcoalg led by @tle_96 openreview.net/pdf?id=RQti5I2… Manifold Learning for Cosmic Structures @ ML4PS led by Ana Sofia Uzoy /w @ClaireLamman openreview.net/pdf?id=mMVdTGC… Balancing Fairness and Accuracy in Graph Learning via Fairness-Constrained Rewiring @neur_reps *oral presentation* led by Jason Wang /w Lukas Fesser openreview.net/pdf?id=msfmcaj… Curvature Estimation on Data Manifolds via Diffusion-augmented Sampling @neur_reps led by Jason Wang /w @bobak_kiani openreview.net/pdf?id=zu24PDR… Data Augmentation: A Fourier Analysis Perspective @neur_reps led by @B_Tahmasebi /w @StefanieJegelka openreview.net/pdf?id=RwPRMX3… Bispectral OT: Dataset Comparison using Symmetry-Aware Optimal Transport @neur_reps led by Annabel Ma /w @kaiyinghou @elmelis openreview.net/pdf?id=5XvfFi4… On a Geometry of Interbrain Networks @neur_reps led by @nico_hinrichs /w Noah Guzman openreview.net/pdf?id=ouNpUPd… Priors in Time: A Generative View of Sparse Autoencoders for Sequential Representations @CogInterp led by @EkdeepL @can_rager @sumedh_hrs openreview.net/pdf?id=WjDtKpt…

Interested in the mathematics of computation in brains and machines? Join us #NeurIPS for the @neur_reps workshop! 🎙️Speakers: Yue Song @tegmark @s_y_chung @SuryaGanguli Razvan Pascanu @kfrankelab Thx to @hopfbifurcator @simoneazeglio @ChasevdG @cashewmake2 @naturecomputes!

Presenting our NeurIPS poster today (11am, #4111): a unified view of saddle-to-saddle dynamics and how neural nets learn modular addition With @giovannimarchet & @FCHEN_AI Dhruva Karkada Jamie Simon Mike DeWeese @SuryaGanguli @ninamiolane Paper: arxiv.org/abs/2506.06489

We are thrilled to see everyone this Sunday at NeurReps! Come join us!

🎉 The challenge is over — and the results are inspiring! 👏 Big shoutout to all winners & contributors! 🚀 More events, challenges & collabs are on the way. Got ideas? DM us and help shape the future of the TDL ecosystem!

The results are finally here! 🤩 A massive shoutout to the winning teams—your work is redefining what’s possible in TDL. Also, a huge thank you to @ARLQ_AI for supporting this challenge. See who took home the gold (and the internships!) in the thread below. 🏆🧵

A final thank you to @ARLQ_AI, @TAGinDS organizing team, reviewers, and every single participant. You have truly expanded the data landscape of Topological Deep Learning. Led with @gbg1441 @lev_telyatnikov ! Explore the new datasets now: github.com/geometric-inte…

The Science Continues 📄 All participants with qualifying submissions will be co-authors on the upcoming white paper in the Topology, Algebra, and Geometry in Data Science PMLR proceedings. 📩 We will be reaching out to all participants soon with more details on the next steps!

🥁 The results are in! We are thrilled to announce the winners of the @TAGinDS Topological Deep Learning Challenge 2025 sponsored by @ARLQ_AI! 🏆 Over the last 2 months, we’ve gotten 44 submissions from 12+ institutions. A huge thank you to the participants and reviewers! 🧵👇