arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2312.09852cs.LGstat.ML
收藏

利用自由形式流学习流形上的分布

Learning Distributions on Manifolds with Free-Form Flows

Heidelberg University(海德堡大学)

Peter Sorrenson, Felix Draxler, Armand Rousselot, Sander Hummerich, Ullrich Köthe

AI总结:

本文提出流形自由形式流(M-FFF),通过将自由形式流框架适配到黎曼流形,实现单次函数求值采样,在匹配或超越特定流形单步方法的同时,比多步方法快约两个数量级。

摘要:

我们提出流形自由形式流(M-FFF),一种用于流形上数据的简单新型生成模型。现有的在任意流形上学习分布的方法在推理时开销很大,因为采样需要求解微分方程。我们的方法通过单次函数求值进行采样,克服了这一限制。关键创新在于通过在流形上最大化似然来优化神经网络,这通过将自由形式流框架适配到黎曼流形而成为可能。M-FFF 可直接适配任何具有已知投影的流形。它始终匹配或优于以往针对特定流形专门设计的单步方法。它通常比基于扩散或流匹配的多步方法快两个数量级,并在若干实验中取得了更好的似然值。我们在 https://github.com/vislearn/FFF 提供代码。

英文摘要:

We propose Manifold Free-Form Flows (M-FFF), a simple new generative model for data on manifolds. The existing approaches to learning a distribution on arbitrary manifolds are expensive at inference time, since sampling requires solving a differential equation. Our method overcomes this limitation by sampling in a single function evaluation. The key innovation is to optimize a neural network via maximum likelihood on the manifold, possible by adapting the free-form flow framework to Riemannian manifolds. M-FFF is straightforwardly adapted to any manifold with a known projection. It consistently matches or outperforms previous single-step methods specialized to specific manifolds. It is typically two orders of magnitude faster than multi-step methods based on diffusion or flow matching, achieving better likelihoods in several experiments. We provide our code at https://github.com/vislearn/FFF.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑