AnchorGK: Anchor-based Incremental and Stratified Graph Learning Framework for Inductive Spatio-Temporal Kriging

2025-12-25

AI Summary

Sparse deployment in sensor networks leads to spatiotemporal observation gaps, and existing methods struggle to jointly address spatial sparsity and heterogeneous availability of auxiliary features across locations. Method: We propose an anchor-based inductive hierarchical graph learning framework. It introduces a novel anchor-driven dynamic hierarchical mechanism to adaptively model both spatial sparsity and feature heterogeneity. A dual-view graph learning layer is designed to jointly aggregate spatial relational structure and multi-source heterogeneous features, while supporting incremental representation updates. The framework integrates graph neural networks, adaptive anchor construction, and hierarchical spatiotemporal graph modeling. Contribution/Results: Our method achieves significant improvements over state-of-the-art approaches on multiple benchmark datasets. It demonstrates strong robustness to incomplete features and unseen locations, substantially enhancing interpolation accuracy and generalization capability.

Technology Category

• Machine Learning: Graph-based Machine Learning
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning
• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems

Application Category

• Graph Algorithms and Modeling for the Web: Algorithms and analysis for heterogeneous, signed, attributed, multi-relational, temporal, higher-order, and annotated Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Spatio-temporal kriging is a fundamental problem in sensor networks, driven by the sparsity of deployed sensors and the resulting missing observations. Although recent approaches model spatial and temporal correlations, they often under-exploit two practical characteristics of real deployments: the sparse spatial distribution of locations and the heterogeneous availability of auxiliary features across locations. To address these challenges, we propose AnchorGK, an Anchor-based Incremental and Stratified Graph Learning framework for inductive spatio-temporal kriging. AnchorGK introduces anchor locations to stratify the data in a principled manner. Anchors are constructed according to feature availability, and strata are then formed around these anchors. This stratification serves two complementary roles. First, it explicitly represents and continuously updates correlations between unobserved regions and surrounding observed locations within a graph learning framework. Second, it enables the systematic use of all available features across strata via an incremental representation mechanism, mitigating feature incompleteness without discarding informative signals. Building on the stratified structure, we design a dual-view graph learning layer that jointly aggregates feature-relevant and location-relevant information, learning stratum-specific representations that support accurate inference under inductive settings. Extensive experiments on multiple benchmark datasets demonstrate that AnchorGK consistently outperforms state-of-the-art baselines for spatio-temporal kriging.

Problem

• Addresses sparse sensor distribution in spatio-temporal kriging
• Handles heterogeneous auxiliary feature availability across locations
• Enables inductive inference for unobserved regions via stratified learning

Innovation

• Anchor-based stratification for graph learning
• Incremental representation to handle incomplete features
• Dual-view graph aggregation for spatio-temporal inference