Abstract:This paper aims to address the challenge of data generation beyond the training data and proposes a framework for Structural Extrapolated Data GEneration (SEDGE) based on suitable assumptions on the underlying data-generating process. We provide conditions under which data satisfying novel specifications can be generated reliably, together with the approximate identifiability of the distribution of such data under certain ``conservative" assumptions, as well as the inherent non-identifiability of this distribution without such assumptions. On the algorithmic side, we develop practical methods to achieve extrapolated data generation, based on a structure-informed optimization strategy or diffusion posterior sampling, respectively. We verify the extrapolation performance on synthetic data and also consider extrapolated image generation as a real-world scenario to illustrate the validity of the proposed framework.
| Subjects: | Machine Learning (cs.LG) |
| Cite as: | arXiv:2604.02482 [cs.LG] |
| (or arXiv:2604.02482v2 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2604.02482 arXiv-issued DOI via DataCite |
From: Jiaqi Sun [view email]
[v1]
Thu, 2 Apr 2026 19:30:24 UTC (17,320 KB)
[v2]
Thu, 14 May 2026 04:48:34 UTC (17,851 KB)
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。