arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2405.17880cs.LG
收藏

扩散拒绝采样

Diffusion Rejection Sampling

KAIST(韩国科学技术院)

Byeonghu Na, Yeongmin Kim, Minsang Park, Donghyeok Shin, Wanmo Kang, Il-Chul Moon

AI总结:

本文提出扩散拒绝采样(DiffRS),通过在每个时间步对齐采样转移核与真实核并拒绝低质量样本,实现更紧的采样误差上界,在基准数据集和快速采样器及文本到图像模型中达到最先进性能。

摘要:

近期在强大的预训练扩散模型方面取得的进展,鼓励开发在训练良好的扩散模型下提高采样性能的方法。本文介绍了扩散拒绝采样(DiffRS),该方法使用一种拒绝采样方案,在每个时间步将采样转移核与真实转移核对齐。所提出的方法可视为一种机制,在每个中间时间步评估样本质量,并根据样本以不同的努力程度对其进行细化。理论分析表明,与预训练模型相比,DiffRS 可以实现更严格的采样误差上界。实证结果展示了 DiffRS 在基准数据集上的最先进性能,以及 DiffRS 在快速扩散采样器和大规模文本到图像扩散模型中的有效性。我们的代码可在 https://github.com/aailabkaist/DiffRS 获取。

英文摘要:

Recent advances in powerful pre-trained diffusion models encourage the development of methods to improve the sampling performance under well-trained diffusion models. This paper introduces Diffusion Rejection Sampling (DiffRS), which uses a rejection sampling scheme that aligns the sampling transition kernels with the true ones at each timestep. The proposed method can be viewed as a mechanism that evaluates the quality of samples at each intermediate timestep and refines them with varying effort depending on the sample. Theoretical analysis shows that DiffRS can achieve a tighter bound on sampling error compared to pre-trained models. Empirical results demonstrate the state-of-the-art performance of DiffRS on the benchmark datasets and the effectiveness of DiffRS for fast diffusion samplers and large-scale text-to-image diffusion models. Our code is available at https://github.com/aailabkaist/DiffRS.

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。

↑