Learning Compositional Sparse Gaussian Processes with a Shrinkage Prior
Learning Compositional Sparse Gaussian Processes with a Shrinkage Prior
复制标题
DOI：
10.1609/aaai.v35i11.17190
复制
发表时间：
2020-12
期刊：
影响因子：
--
通讯作者：
Anh Tong;Toan Tran;H. Bui;Jaesik Choi
中科院分区：
文献类型：
--
作者：
Anh Tong;Toan Tran;H. Bui;Jaesik Choi
在高斯过程（GP）模型的学习中，核函数的选择是一个重要的问题，因为每个核结构具有不同的模型复杂度和数据适应度。最近，自动核组合方法不仅提供了准确的预测，但也有吸引力的解释，通过基于搜索的方法。然而，现有的方法遭受缓慢的内核组成学习。为了解决大规模的数据，我们提出了一个新的稀疏近似后验GP，MultiSVGP，构造从组的诱导点与个人添加剂内核组成的内核。我们证明，这种近似提供了一个更好的适合学习组成内核的经验观察。与传统的稀疏GP相比，我们还提供了理论上的误差界的理由。与基于搜索的方法相比，我们提出了一种新的概率算法来学习的核心组成，通过处理的稀疏性与马蹄事先的内核选择。我们证明了我们的模型可以捕获时间序列的特征，大大减少了计算时间，并在真实世界的数据集上具有竞争力的回归性能。
Choosing a proper set of kernel functions is an important problem in learning Gaussian Process (GP) models since each kernel structure has different model complexity and data fitness. Recently, automatic kernel composition methods provide not only accurate prediction but also attractive interpretability through search-based methods. However, existing methods suffer from slow kernel composition learning. To tackle large-scaled data, we propose a new sparse approximate posterior for GPs, MultiSVGP, constructed from groups of inducing points associated with individual additive kernels in compositional kernels. We demonstrate that this approximation provides a better fit to learn compositional kernels given empirical observations. We also provide theoretically justification on error bound when compared to the traditional sparse GP. In contrast to the search-based approach, we present a novel probabilistic algorithm to learn a kernel composition by handling the sparsity in the kernel selection with Horseshoe prior. We demonstrate that our model can capture characteristics of time series with significant reductions in computational time and have competitive regression performance on real-world data sets.