Research map / Graph learning / GNN architectures
Graph Neural Networks for Learning Equivariant Representations of Neural Networks
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
The paper proposes equivariant graph neural networks that process neural network parameters by representing weight spaces as graphs, achieving permutation equivariance without intricate weight-sharing schemes.
Problem
Neural networks that process other networks' parameters should respect the permutation symmetry of hidden units, but existing approaches either ignore it or rely on complex weight-sharing patterns to achieve equivariance.
Approach
Represent a neural network's parameters as a graph whose nodes are neurons and edges carry weights, and apply permutation-equivariant graph neural networks to this representation, so that permutation symmetry is built in directly.
Experimental setup
Metrics: accuracy
Baselines: DWSNets NFN
Builds on: message passing neural networks DeepSets
Closest papers by meaning
- Graph Metanetworks for Processing Diverse Neural Architectures (ICLR 2024 · similarity 0.91)
- Are High-Degree Representations Really Unnecessary in Equivariant Graph Neural Networks? (NeurIPS 2024 · similarity 0.90)
- Permutation Equivariant Neural Controlled Differential Equations for Dynamic Graph Representation Learning (NeurIPS 2025 · similarity 0.89)
- A Characterization Theorem for Equivariant Networks with Point-wise Activations (ICLR 2024 · similarity 0.89)
- Equivariant Machine Learning on Graphs with Nonlinear Spectral Filters (NeurIPS 2024 · similarity 0.89)
- Scale Equivariant Graph Metanetworks (NeurIPS 2024 · similarity 0.89)
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.