🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

Un évènement sur l'IA où je participais a été chamboulé par Extinction Rébellion (mauvaise cible: ça parlait des limites de l'IA). C'est symptomatique du contexte de communication contrôlée ne laissant aux voix dissonantes uniquement une place violente.

I'm thrilled to announce that my #ERCStG project has been accepted 🤓 **MALAGA: Reinventing the Theory of Machine Learning on Large Graphs** Many job openings coming up, see nkeriven.github.io/malaga for updates! Thank you European Research Council (ERC) and all my collaborators past and future

🤗Officially started Ph.D. with Ievgen Redko, Romain Tavenard and Laetitia Chapel Inria IRISA on Transformers & Distribution Shifts 🥳🇨🇦 Also, 2 papers accepted at #NeurIPS2024 📈 *Spotlight* arxiv.org/pdf/2406.10327 ✋🏾 MaNO arxiv.org/pdf/2405.18979 More details soon!

Large Language Models as Markov Chains ift.tt/uPbQTkA

Stein's Lemma states that for a normally distributed variable X, the expected value E[Xg(X)] = E[g’(X)] for any g absolutely continuous (derivative a.e.) such that E[|g’(X)|] < ∞. It is a central result for characterizing Gaussian data arxiv.org/abs/2104.08302

A simple, yet overlooked idea: LLMs with a finite vocabulary and context window are (finite) Markov chains :) An 🤩 internship of @oussamazekri_in collaboration with now officially our 1st year Ph.D. student Ambroise Odonnat Abdelhakim Benechehab @bleistein_linus & N. Boullé A 🧵⬇️

🚨 Interested in domain adaptation and generalization for 3D data? 🚨 It’s tough to keep up with all the new amazing work. That’s why I’ve curated a comprehensive, easy-to-navigate list of publications. 📚 👉 Repo: bit.ly/DA_DG_3D

🏆Didn't get the Physics Nobel Prize this year, but really excited to share that I've been named one of the #FWIS2024 Fondation L'Oréal-UNESCO 🏛️ #Education #Sciences #Culture 🇺🇳 French Young Talents alongside 34 amazing young researchers! This award recognizes my research on deep learning theory #WomenInScience 👩‍💻

Congratulations to Dr. Dagréou Mathieu Dagréou for a brillant PhD defense! Moreau Thomas, Pierre Ablin and I were lucky to have you as a student.

Bilevel optimization problems with multiple inner solutions come typically in two flavors: optimistic and pessimistic. Optimistic assumes the inner problem selects the best solution for the outer objective, while pessimistic assumes the worst-case solution is chosen.

Brenier's theorem states that the optimal transport map between two probability measures for quadratic cost is the gradient of a convex function. Moreover, it is uniquely defined up to a Lebesgue negligible set. ceremade.dauphine.fr/~carlier/Breni…

Cool work on a differentiable variant of dynamic time warping by Romain Tavenard rtavenar.github.io/blog/softdtw.h…

A Monge map, i.e., a solution to optimal transport Monge problems, may not always exist, be unique, or be symmetric with respect to the source and target distributions. It was one of the motivation to introduce Kantorovich relaxation. math.univ-lyon1.fr/~santambrogio/…

I have several internship/PhD positions on Graph ML available with ERC MALAGA 🤓 See details here: nkeriven.github.io/malaga/ Don't hesitate to contact me! (same if you are interested in a post-doc related to graph ML)

I have multiple openings for M2 internship / PhD / postdoc in Nice (France) on topics related to bilevel optimization, automatic differentiation and safe machine learning. More details on my webpage samuelvaiter.com Contact me by email, and feel free to forward/RT :)

Our Transactions on Machine Learning Research paper "Gradient scarcity with Bilevel Optimization for Graph Learning" (w/ H Ghanem, Samuel Vaiter ) was accepted as an oral presentation at Learning on Graphs Conference 2025 🤓 100% free and online, come check it out! arxiv.org/abs/2303.13964

We have an accepted paper at NeurIPS 2024 about supervised graph prediction with transformers and optimal transport loss with Paul Krzakala, Junjie Yang, Florence d’Alché-Buc, Charlotte Laclau and Matthieu Labeau. Details in 🧵 arxiv.org/pdf/2402.12269

Come work with us on tslearn in beautiful Rennes! (deadline for application is soon!) jobs.inria.fr/public/classic…

📣 New preprint 📣 **Differentiable Generalized Sliced Wasserstein Plans** w/ L. Chapel Romain Tavenard We propose a Generalized Sliced Wasserstein method that provides an approximated transport plan and which admits a differentiable approximation. arxiv.org/abs/2505.22049 1/5