PulseAugur
EN
LIVE 06:49:57

Paper details mathematical principles of generative modeling techniques

This paper explores the mathematical underpinnings of generative modeling, drawing connections between optimal transport and techniques like Schrödinger bridge and flow matching. It aims to provide a high-level overview of these principles for those interested in the field. AI

RANK_REASON The item is an arXiv preprint detailing mathematical principles of generative modeling. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv stat.ML →

AI-generated summary · Google Gemini · from 3 sources. How we write summaries →

Paper details mathematical principles of generative modeling techniques

COVERAGE [3]

  1. arXiv stat.ML TIER_1 English(EN) · Titouan Vayer (COMPACT) ·

    Notes on generative modeling: flow matching, diffusion, optimal transport and Schr{\"o}dinger bridge

    arXiv:2606.30053v1 Announce Type: new Abstract: These notes recapitulate the high level mathematical principles behind different techniques for generative modeling. I show the connections between optimal transport and standard techniques such as Schr{\"o}dinger bridge and flow ma…

  2. arXiv stat.ML TIER_1 English(EN) · Ziyu Chen, Markos A. Katsoulakis, Benjamin J. Zhang ·

    Robustness and Structure Preservation in Flow-Based Generative Models via Wasserstein Path-Space Divergences

    arXiv:2410.01244v2 Announce Type: replace Abstract: We introduce a novel Wasserstein-1 ($W_1$) path-space divergence for stochastic and deterministic dynamics and establish a Wasserstein Uncertainty Propagation (WUP) theorem that bounds the $W_1$ distance between terminal distrib…

  3. arXiv stat.ML TIER_1 English(EN) · Titouan Vayer ·

    Notes on generative modeling: flow matching, diffusion, optimal transport and Schr{ö}dinger bridge

    These notes recapitulate the high level mathematical principles behind different techniques for generative modeling. I show the connections between optimal transport and standard techniques such as Schr{ö}dinger bridge and flow matching.