ADGC – Awesome Deep Graph Clustering

What the repo is

ADGC is a curated bibliography of research on deep graph clustering – a sub‑field of graph neural networks that tries to automatically group the nodes of a graph into meaningful communities. The repository does not contain a single codebase; instead it lists papers, links to the original PDFs, and (when available) links to the authors’ implementations.

Why it matters

Graph clustering is used in many real‑world problems: social‑network community detection, protein‑interaction analysis, recommendation systems, and single‑cell genomics. Recent “deep” approaches combine graph neural networks (GNNs) with clustering objectives, often improving over classic spectral or modularity‑based methods. ADGC helps researchers keep up with the rapid pace of new models, benchmarks, and datasets.

How the repo is organised

Section	Content
Survey papers	Key overview articles that summarise the field and provide taxonomy.
LLM‑based Deep Graph Clustering	Works that incorporate large language models to guide or augment clustering.
New‑architecture Deep Graph Clustering	Papers proposing novel GNN designs, reinforcement‑learning tricks, hyperbolic embeddings, etc.
Temporal Deep Graph Clustering	Methods that handle evolving graphs over time.
Clustering with Unknown Number of Clusters	Approaches that do not require a pre‑specified k.
Reconstructive Deep Graph Clustering	Auto‑encoder‑style models that reconstruct graph structure while clustering.
Tables	Each entry gives year, title, venue, a link to the paper, and (when present) a link to the code repository.

How to use it

1. Find a topic – e.g., “temporal clustering” or “LLM‑guided clustering”.
2. Browse the table for the most recent papers (the list is kept up‑to‑date with 2026 entries).
3. Follow the paper link to read the PDF; if a code link is provided, you can clone the implementation directly from the author’s repo.
4. Cite – the README asks users to cite 2‑3 of the listed papers when re‑using the datasets or code.

Who should look at this

• Graduate students or post‑docs starting a project on graph clustering.
• Practitioners needing the latest open‑source implementations for community detection in their data pipelines.
• Reviewers or survey writers who want a ready‑made list of recent contributions.

---

All information above is taken directly from the repository’s README; the repo itself does not host any executable software.