arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2211.08036stat.MLcs.LGstat.ME
收藏

高斯过程的可证可靠大规模抽样

Provably Reliable Large-Scale Sampling from Gaussian Processes

Bristol University(布里斯托尔大学);IBM Research(IBM研究院)

Anthony Stephenson, Robert Allison, Edward Pyzer-Knapp

摘要:

当比较近似高斯过程(GP)模型时,能够从任何GP生成数据可能是有帮助的。如果我们对近似方法在规模上的表现感兴趣,我们可能希望生成非常大的合成数据集来评估它们。简单地这样做会消耗\(\mathcal{O}(n^3)\)触发器和\(\mathcal{O}(n^2)\)内存来生成大小为\(n\)的样本。我们演示了如何将这样的数据生成扩展到大的\(n\),同时仍然提供保证,以高概率,样本是无法区分的样本从所需的GP。

英文摘要:

When comparing approximate Gaussian process (GP) models, it can be helpful to be able to generate data from any GP. If we are interested in how approximate methods perform at scale, we may wish to generate very large synthetic datasets to evaluate them. Naïvely doing so would cost \(\mathcal{O}(n^3)\) flops and \(\mathcal{O}(n^2)\) memory to generate a size \(n\) sample. We demonstrate how to scale such data generation to large \(n\) whilst still providing guarantees that, with high probability, the sample is indistinguishable from a sample from the desired GP.

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。

↑