A PyTorch implementation of "SimGNN: A Neural Network Approach to Fast Graph Similarity Computation" (WSDM 2019).
-
Updated
Jan 12, 2023 - Python
A PyTorch implementation of "SimGNN: A Neural Network Approach to Fast Graph Similarity Computation" (WSDM 2019).
A python package for graph kernels, graph edit distances, and graph pre-image problem.
A PyTorch Geometric implementation of SimGNN with some extensions.
Code for "Multilevel Graph Matching Networks for Deep Graph Similarity Learning"
Quantifying Pairwise Chemical Similarity for Polymers
[ICCVw 2023] SSIG: A Visually-Guided Graph Edit Distance for Floor Plan Similarity
A Python/Cython package for graph edit distances and graph matching
[AAAI 2025] GraSP: Simple yet Effective Graph Similarity Predictions
A Jupyter notebook for a project centered around 'Group Recommendation Systems (GRS)' utilizing the 'GcPp' clustering approach.
Calculating Pairwise Similarity of Polymer Ensembles via Earth Mover’s Distance
Code and Data for the paper: "Gelato: Graph Edit Distance via Autoregressive Neural Combinatorial Optimization" (ICLR 2026)
DRESS: A Continuous Framework for Structural Graph Refinement
WP 1.4 — the first Mesoamerican schema (Popol Vuh), with zero Old World contact, lands squarely in the branching-tree family. Geographic generality confirmed. Adds edge-weighted GED. 17 traditions.
WP 1.5 — 22 traditions, and the first intra-tradition split: the Rig Veda yields both a linear chain and a branching tree, showing topology tracks cosmological structure rather than culture or geography.
WP 1.6 — 25 traditions and a five-way zero-contact convergence: Tibetan, Greek, Chinese, Vedic and Orphic schemas all produce identical 5-node depth-4 chains. Adds a graduated 9x9 role-similarity matrix.
WP 1.2 — adding Proclus and Suhrawardi takes the emanation-topology corpus to 11 traditions and 5 exact structural isomorphisms. Fully reproducible pipeline, ~115s runtime. Zenodo 10.5281/zenodo.19305988.
Graph-theoretic comparison of 9 cosmological hierarchies encoded as labelled DAGs — VF2 isomorphism, Weisfeiler-Leman kernels, graph edit distance, 7,000 null models. Two exact cross-cultural isomorphisms.
WP 1.3 — Zoroastrian, Manichaean and Orphic schemas take the corpus to 14 traditions. Family separation ratio rises to 2.87x; 8 exact isomorphisms across 4 clusters. Adds a succession edge type.
Build a supergraph of a dataset
A GNN-based framework to derive Graph Edit Distance edit operations
To associate your repository with the graph-edit-distance topic, visit your repo's landing page and select "manage topics."