arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2110.09807stat.MLcs.LGcs.SIeess.SP
收藏

学习如何学习图拓扑

Learning to Learn Graph Topologies

University of Oxford(牛津大学);Shanghai Jiao Tong University(上海交通大学)

Xingyue Pu, Tianyue Cao, Xiaoyun Zhang, Xiaowen Dong, Siheng Chen

AI总结:

针对经典图拓扑学习方法依赖显式凸先验函数、灵活性不足的问题,提出基于学习优化思想的端到端模型,将迭代算法展开为神经网络并用变分自编码器优化拓扑,效率更优。

摘要:

学习图拓扑以揭示数据实体间的潜在关系,在各类机器学习与数据分析任务中发挥着重要作用。在结构化数据于图上平滑变化的假设下,该问题可被建模为半正定锥上的正则化凸优化问题,并通过迭代算法求解。经典方法需要借助显式凸函数来体现通用拓扑先验,例如用于保证稀疏性的$\ell_1$惩罚项,这限制了学习丰富拓扑结构时的灵活性与表达能力。我们基于学习优化(L2O)的思想,提出学习一种从节点数据到图结构的映射方法。具体而言,我们的模型首先将迭代原始-对偶分裂算法展开为神经网络。其中关键的结构近端投影被替换为变分自编码器,用于优化估计得到的图,提升其拓扑特性。该模型以端到端的方式,利用节点数据与图样本对进行训练。在合成数据与真实数据上的实验表明,在学习具有特定拓扑性质的图时,我们的模型比经典迭代算法效率更高。

英文摘要:

Learning a graph topology to reveal the underlying relationship between data entities plays an important role in various machine learning and data analysis tasks. Under the assumption that structured data vary smoothly over a graph, the problem can be formulated as a regularised convex optimisation over a positive semidefinite cone and solved by iterative algorithms. Classic methods require an explicit convex function to reflect generic topological priors, e.g. the $\ell_1$ penalty for enforcing sparsity, which limits the flexibility and expressiveness in learning rich topological structures. We propose to learn a mapping from node data to the graph structure based on the idea of learning to optimise (L2O). Specifically, our model first unrolls an iterative primal-dual splitting algorithm into a neural network. The key structural proximal projection is replaced with a variational autoencoder that refines the estimated graph with enhanced topological properties. The model is trained in an end-to-end fashion with pairs of node data and graph samples. Experiments on both synthetic and real-world data demonstrate that our model is more efficient than classic iterative algorithms in learning a graph with specific topological properties.

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。

↑