跳到正文
研踪PaperTrail
← 研踪 PaperTrail

arXiv:2605.09439 · v2

Inverse Design for Conditional Distribution Matching

首次提交 2026-05-10 17:27
版本更新 2026-11-18 03:08
雷达观测 2026-11-19 19:08
arXiv ↗PDF ↗

摘要

Generative models are powerful tools for sampling from a learned distribution P(Y∣X)\mathcal{P}(Y \mid X), and inverse-design methods invert this map to find an input xx that produces a desired point output y∗y^*. However, many design goals are naturally distributional rather than pointwise, incorporating the inherent uncertainty of YY and targeting a specific form for it, a task not addressed by standard inverse design. To address this issue we introduce Conditional Distribution Matching (CDM), a new inverse-design problem class in generative modeling: given a joint distribution P(X,Y)\mathcal{P}(X, Y) and a target distribution G(Y)\mathcal{G}(Y), find an input x∗x^* whose induced conditional distribution P(Y∣X=x∗)\mathcal{P}(Y \mid X = x^*) matches G\mathcal{G}. We formally define two variants: Conditional Distribution Matching Sampling (CDMS) and Conditional Distribution Matching Optimization (CDMO). To solve these problems, we propose MLGD-F (Matching-Loss Guided Diffusion with a Fast inner sampler), a plug-and-play inference-time algorithm that combines a pretrained score-based diffusion model with a pretrained fast conditional sampler, requiring no additional training or fine-tuning. By leveraging single-step conditional sampling, MLGD-F enables tractable gradient computation, making the estimation of P(Y∣X)\mathcal{P}(Y \mid X) both memory-efficient and computationally lightweight. We validate MLGD-F on synthetic benchmarks, structured image transformations, and generative editing optimization, demonstrating reliable recovery of inputs whose conditional distributions match diverse user-specified targets, including discrete mixtures and continuous low-rank supports.

AI 论文解读

基于摘要 · 非全文审阅

此版本尚未生成 AI 解读。解读由后台任务生成并按论文版本保存,登录后共享已有结果。

版本差异

本站首次收录为 v2,暂无更早版本可对照。

观测到的版本

相邻记录可能跳过中间版本;首次发现 v2 及以上也会标为版本更新。

v2 · 2026-11-18 03:08 · 最新观测版本

Inverse Design for Conditional Distribution Matching

Generative models are powerful tools for sampling from a learned distribution P(Y∣X)\mathcal{P}(Y \mid X), and inverse-design methods invert this map to find an input xx that produces a desired point output y∗y^*. However, many design goals are naturally distributional rather than pointwise, incorporating the inherent uncertainty of YY and targeting a specific form for it, a task not addressed by standard inverse design. To address this issue we introduce Conditional Distribution Matching (CDM), a new inverse-design problem class in generative modeling: given a joint distribution P(X,Y)\mathcal{P}(X, Y) and a target distribution G(Y)\mathcal{G}(Y), find an input x∗x^* whose induced conditional distribution P(Y∣X=x∗)\mathcal{P}(Y \mid X = x^*) matches G\mathcal{G}. We formally define two variants: Conditional Distribution Matching Sampling (CDMS) and Conditional Distribution Matching Optimization (CDMO). To solve these problems, we propose MLGD-F (Matching-Loss Guided Diffusion with a Fast inner sampler), a plug-and-play inference-time algorithm that combines a pretrained score-based diffusion model with a pretrained fast conditional sampler, requiring no additional training or fine-tuning. By leveraging single-step conditional sampling, MLGD-F enables tractable gradient computation, making the estimation of P(Y∣X)\mathcal{P}(Y \mid X) both memory-efficient and computationally lightweight. We validate MLGD-F on synthetic benchmarks, structured image transformations, and generative editing optimization, demonstrating reliable recovery of inputs whose conditional distributions match diverse user-specified targets, including discrete mixtures and continuous low-rank supports.

查看此版本解读 →查看此版本原文 ↗