arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2210.01549cs.LGcs.SIstat.ML
收藏

图扩散模型受益于离散状态空间

Diffusion Models for Graphs Benefit From Discrete State Spaces

ETH Zurich(苏黎世联邦理工学院);Swiss Data Science Center(瑞士数据科学中心)

Kilian Konstantin Haefeli, Karolis Martinkus, Nathanaël Perraudin, Roger Wattenhofer

AI总结:

针对图生成中连续高斯扩散破坏离散结构的问题,本文提出离散状态空间的前向马尔可夫加噪过程,在四个数据集和多种架构上提升样本质量并将采样加速30倍。

摘要:

去噪扩散概率模型和分数匹配模型已被证明在生成任务中非常强大。尽管这些方法也已被应用于离散图的生成,但迄今为止它们都依赖于连续的高斯扰动。与此不同,本研究建议在前向马尔可夫过程中使用离散噪声。这确保了在每个中间步骤中图都保持离散。与先前方法相比,我们在四个数据集和多种架构上的实验结果表明,使用离散加噪过程能够产生质量更高的生成样本,其表现为平均 MMD 降低至原来的 1/1.5。此外,去噪步数从 1000 步减少到 32 步,使采样过程加快 30 倍。

英文摘要:

Denoising diffusion probabilistic models and score-matching models have proven to be very powerful for generative tasks. While these approaches have also been applied to the generation of discrete graphs, they have, so far, relied on continuous Gaussian perturbations. Instead, in this work, we suggest using discrete noise for the forward Markov process. This ensures that in every intermediate step the graph remains discrete. Compared to the previous approach, our experimental results on four datasets and multiple architectures show that using a discrete noising process results in higher quality generated samples indicated with an average MMDs reduced by a factor of 1.5. Furthermore, the number of denoising steps is reduced from 1000 to 32 steps, leading to a 30 times faster sampling procedure.

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。

↑