Graph Convolutional Network with Multi-View Topology for Lightweight Skeleton-Based Action Recognition Skeleton-based action recognition is an important subject in deep learning. Graph Convolutional Networks (GCNs) have demonstrated strong performance by modeling the human skeleton as a natural topological graph, representing the connections between joints. However, most existing methods rely on non-adaptive topologies or insufficiently expressive representations. To address these limitations, we propose a Multi-view Topology Refinement Graph Convolutional Network (MTR-GCN), which is efficient, lightweight, and delivers high performance. Specifically: We propose a new spatial topology modeling approach that incorporates two views. A dynamic view fuses joint information from dual streams in a pairwise manner, while a static view encodes the shortest static paths between joints, preserving the original connectivity relationships. We propose a new MultiScale Temporal Convolutional Network...
Network science disentangles internal climate variability in global spatial dependence structures A comprehensive characterization of internal climate variability (ICV) in initial-condition (IC) large ensembles of Earth system models (ESMs) remains a significant challenge in climate science. In this study, we leverage the spatial connectivity structures of temperature networks to characterize ICV, observing substantial differences across ensemble members, particularly in the prevalence of long-range connections. Based on this feature, we introduce the ‘Connectivity Ratio’ (CR), a new quantifier that captures long-range spatial connectivity within climate networks. CR is applied to two ESMs, EC-Earth3 and MPI-ESM1-2-LR, to evaluate structural variability across IC ensemble members, models, and climate time horizons. CR reveals systematic differences in long-range connectivity between forced and unforced simulations, as well as across future climate periods. As such, CR provides an inter...