domain-graph
Skill8 starsChanged 4 months agoMITView on GitHub
Use for graph machine learning: node classification, link prediction, graph classification or regression, recommendation on graphs, molecule or network analysis, community detection, and graph anomaly detection with graph neural networks. Picks the GNN by task, enforces transductive vs inductive splits (a common leakage source), and uses graph metrics. Triggers on 'graph neural network', 'GNN', 'node classification', 'link prediction', 'knowledge graph', 'molecule', 'social network', 'GCN', 'GraphSAGE', 'GAT'.
In this repository22 other files it gives its agents.
- accuracy-improvement-loopSkill
- data-rigor-and-leakageSkill
- dataset-profilingSkill
- deployment-explainabilitySkill
- domain-3dSkill
- domain-anomaly-detectionSkill
- domain-audio-speechSkill
- domain-computer-visionSkill
- domain-generativeSkill
- domain-medical-imagingSkill
- domain-multimodalSkill
- domain-nlp-llmSkill
- domain-recommenderSkill
- domain-reinforcement-learningSkill
- domain-tabularSkill
- domain-time-seriesSkill
- domain-videoSkill
- literature-reviewSkill
- ml-research-methodologySkill
- notebook-deliverySkill
- rigorous-evaluationSkill
- training-optimizationSkill
The file
We split this skill into the rules it gives an agent
1 rule. About 0 are common, 1 is distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 8
Use PyTorch Geometric (PyG) or DGL.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Graph ML - Method Selection
Use PyTorch Geometric (PyG) or DGL. GNNs need node features, so add structural features (degree, node2vec) when raw features are weak.
Decision table
| Task | Recommended | Notes |
| Node classification, static graph | GCN (simple, strong baseline), then GAT | transductive; GCN is a remarkably strong baseline. |
| Large or industrial, or unseen nodes | GraphSAGE (inductive, neighbor sampling) | generalizes to nodes and graphs not seen in training. |
| Graph classification or regression | GIN (maximally expressive) with global pooling | molecules, program graphs. |
| Link prediction or graph recommendation | GraphSAGE or LightGCN with negative sampling | metrics are AUC or AP, and Hits@K or MRR. |
| Weak or missing features | add structural or node2vec features | GNNs underperform without informative features. |
Cross-cutting practice
- Leakage (critical): be explicit about transductive (mask nodes within one graph) versus inductive (held-out nodes or graphs) evaluation. For link prediction, remove the test edges from the message-passing graph during training, otherwise the model sees the answer. For graph classification, split by graph.
- Metrics: node and graph classification use accuracy or macro F1 (macro when imbalanced); link prediction uses AUC or AP, Hits@K, MRR.
- Explainability: GNNExplainer or integrated gradients (which nodes and edges mattered), and attention weights for GAT.
- Recent options (2022-2024): Graph Transformers (Graphormer, GraphGPS, Exphormer). A 2024 reassessment shows well-tuned classic GNNs (GCN, GAT, GraphSAGE) match or beat them on most benchmarks, so tune the simple baseline before reaching for a transformer.
- Improve accuracy: use
accuracy-improvement-loop; evaluate withrigorous-evaluation.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.