Research map / Graph learning / Graph self-supervised learning
GEM: A Dual-Scale Architecture for Graph-Level Hierarchical Representation Learning
Zhaowei Wu, Kaizhong Zheng, Guangmingzi Yang, Shuai Jiang, Liangjun Chen, Badong Chen
Read the paperCheck your idea against 137,217 papers
AI-extracted summary. Written by AI from the paper's abstract only. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Proposes Graph Encoding Mamba (GEM), a dual-scale architecture for graph-level representation learning that aims to capture local clustering and global propagation in small-world networks.
Problem
Graph-level representation learning must capture both dense local clustering and efficient long-range propagation in small-world networks; GNNs suffer over-smoothing for long-range modeling, while sequence/state-space models lack inductive bias for graph topology.
Approach
Combines graph structure inductive bias with a Mamba state-space backbone in a dual-scale design (local and global), intended to reconcile local clustering with global propagation without GNN over-smoothing. Details beyond the abstract are not available.
Experimental setup
Builds on: Mamba
Closest papers by meaning
- Dual Mamba for Node-Specific Representation Learning: Tackling Over-Smoothing with Selective State Space Modeling (AAAI 2026 · similarity 0.90)
- Accurate Learning of Graph Representations with Graph Multiset Pooling (ICLR 2021 · similarity 0.89)
- Hierarchical Graph Representation Learning with Pooling-Induced Substructures (NeurIPS 2026 · similarity 0.89)
- Distance Encoding: Design Provably More Powerful Neural Networks for Graph Representation Learning (NeurIPS 2020 · similarity 0.88)
- Translating Subgraphs to Nodes Makes Simple GNNs Strong and Efficient for Subgraph Representation Learning (ICML 2024 · similarity 0.87)
- Learning Representations for Hierarchies with Minimal Support (NeurIPS 2024 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
