🔥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

Sparse solutions of the kernel herding algorithm by improved gradient approximation. (arXiv:2105.07900v2 [math.NA] UPDATED) ift.tt/xCyT98V

In our preprint, for the 1st time we make neural operators equivariant with respect to PDE symmetry groups. These can be very complicated, and often only the Lie algebra is known, so a universal method is needed - a 🧵. 🔗 Read the full paper here: arxiv.org/abs/2410.02698

젊을때 만이 못 놀아본 사람은 많이 놀아보라하고 젊을때 노느라 공부 저축 안한 사람은 젊을때 열심히 살라하고 인간은 어떤 삶을 살아도 후회하고 자기가 느끼는 결핍만 줄줄이 읊기에 여념이 없는듯

Koopman Operator Theory provides a framework for #metalearning of #neuralnetworks . Our paper arxiv.org/abs/2302.09160 - accepted as #NeurIPS highlight - with Will et al. provides methodology for comparing neural network training protocols. #AI #ML #KoopmanOperator

내가 박사만 따면 박사과정 때려친다

A cat roars while looking at its reflection in the mirror but instead sees itself as a lion roaring #veo2

A Diagrammatic Approach to Improve Computational Efficiency in Group Equivariant Neural Networks ift.tt/oCt9IDm

I don't know how to raise kids with the AI Shepard tone in the background.

Extended Symmetry and Geometry in Neural Representations abstract of my full paper on neural network symmetrisation in Markov categories: arxiv.org/abs/2412.09469 See for an overview of the story in terms of deterministic functions and Markov kernels rather than general Markov categories.

Why do diffusion models generalise at all? It's not obvious that they would. It turns out underfitting plays an important role, as well as the architectural inductive biases of locality and translation equivariance. What other kinds of symmetry and structure could we hardcode? 🤔

Little is known about how deep networks interact with structure in data. An important aspect of this structure is symmetry (e.g., pose transformations). Here, we (w/ Stéphane Deny) study the generalization ability of deep networks on symmetric datasets: arxiv.org/abs/2412.11521

A proof is a program, and LLMs are very good at finding programs. Now, we only need to generate programs in an adequate programming language (lean) in an smarter way than leverages the structure of the program language (type theory/hott?) instead of only plain text.

Variational Flow Matching goes Riemannian! 🔮 In this preliminary work, we derive a variational objective for probability flows 🌀 on manifolds with closed-form geodesics. My dream team: Floor Eijkelboom Alison Erik Bekkers 💥 📜 arxiv.org/abs/2502.12981 🧵1/5

Learning with Exact Invariances in Polynomial Time. arxiv.org/abs/2502.19758

A meta-point of this paper is that category theory has utility for reasoning about current problems of interest in mainstream machine learning. The theory is predictive, not just descriptive. 🧵(1/6)

'The founders of Google computed the Perron-Frobenius eigenvector of the web graph and became billionaires.' -- A. Brouwer and W. Haemers

Ascend to the peak of a haunted kingdom in Hollow Knight: Silksong, arriving on #NintendoSwitch2 this year.

Will be presenting: - a (spotlight!) paper with Jinwoo Kim on RWNNs: arxiv.org/abs/2407.01214 - a workshop (oral!) paper on RG-VFM (Delta Workshop) arxiv.org/abs/2502.12981 - a workshop paper on a Spectral Study of DiGress (Delta & XAI4S Workshops) openreview.net/pdf?id=vPx5855…

Symmetry is the fundamental property of crystals, yet generative models don't yield crystals with realistic symmetries We solved that with SymmCD and can get crystals from any of the 230 space groups Learn more at our #ICLR poster w/Daniel Levy Siba Smarak Panigrahi @ ICLR2025 ✈️🇸🇬 arxiv.org/abs/2502.03638 🧵