arXiv cs.AI· Keshu Wu, Sixu Li, Zihao Li, Zhiwen Fan, Xiaopeng Li, Yang Zhou·· 3 天前AI 评分20
MoHL：用超图拉普拉斯混合模型从不完整时空数据中学习高阶交互
Learning Higher-Order Interactions from Incomplete Spatiotemporal Data with Mixtures of Hypergraph Laplacians
AI 导读
研究者提出 MoHL（mixture of hypergraph Laplacians），可直接从不完整的时空数据中学习分组关系结构，通过从观测到的共变中挖掘重叠传感器组、为每种交互阶构建拉普拉斯分量，并借助模拟缺失模式的留出集学习混合权重。在三个交通网络上的实验显示，长时段数据缺失下高阶分组收益最大，MoHL 在整日缺失场景下取得所评估基线中最低误差，且学习参数量比深度插补模型少数个数量级。
正文 · 原文
Abstract:Traffic sensor data exhibit strong temporal and cross-sensor dependencies, yet long gaps erase both a sensor's local history and the joint observations needed to learn its relationships with other sensors. We introduce MoHL, a mixture of hypergraph Laplacians that learns group-wise relational structure directly from incomplete spatiotemporal data. MoHL discovers overlapping sensor groups from observed co-variation, constructs a Laplacian component for each interaction order, and learns their mixture through holdouts that mimic the missingness pattern. By sharing weights across group-induced couplings, MoHL pools scattered evidence instead of estimating each coupling independently. We characterize the induced weight-sharing structure and analyze the conditions under which sharing reduces estimation risk. Experiments on three traffic networks show that higher-order groups are most beneficial under long gaps, while support-matched comparisons demonstrate the advantage of shared weights when relational evidence is scarce. MoHL achieves the lowest error among the evaluated baseline methods under day-long gaps, with orders of magnitude fewer learned parameters than deep imputers. These results show how group-wise sharing can sustain reliable reconstruction precisely when missingness depletes the evidence needed to learn cross-sensor dependence.
| Subjects: | Machine Learning (cs.LG); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2605.17316 [cs.LG] |
| (or arXiv:2605.17316v2 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2605.17316 arXiv-issued DOI via DataCite |
Submission history
From: Keshu Wu [view email]
[v1]
Sun, 17 May 2026 08:15:16 UTC (649 KB)
[v2]
Sun, 11 Oct 2026 02:28:48 UTC (4,751 KB)
来源：arXiv cs.AI · arxiv.org