GraphUniverse: Enabling Systematic Evaluation of Inductive Generalization

Generate families of graphs with finely controllable properties for systematic evaluation of inductive graph learning models.

Quick Start | Reproduce Validation Experiment | Interactive Demo

Key Features

Current graph learning benchmarks are limited to single-graph, transductive settings. GraphUniverse enables the first systematic evaluation of inductive generalization by generating entire families of graphs with:

• Consistent Semantics: Communities maintain stable identities across graphs
• Fine-grained Control: Tune homophily, degree distributions, community structure
• Scalable Generation: Linear scaling, thousands of graphs per minute
• Validated Framework: Comprehensive parameter sensitivity analysis
• Interactive Tools: Web-based exploration and visualization

Quick Start

Installation

# Install directly from GitHub
pip install git+https://github.com/LouisVanLangendonck/GraphUniverse.git

# Or clone and install in development mode
git clone https://github.com/LouisVanLangendonck/GraphUniverse.git
cd GraphUniverse
pip install -e .

Basic Usage (see examples/quickstart.py)

Option 1: Via individual classes

from graph_universe import GraphUniverse, GraphFamilyGenerator

# Create universe with detailed parameters
universe = GraphUniverse(K=8, edge_propensity_variance=0.3, feature_dim=10)

# Generate family with full parameter control
family = GraphFamilyGenerator(
    universe=universe,
    min_n_nodes=25, 
    max_n_nodes=50,
    min_communities=2,
    max_communities=7,
    homophily_range=(0.2, 0.8),
    avg_degree_range=(2.0, 10.0),
    degree_distribution="power_law",
    power_law_exponent_range=(2.0, 5.0),
    degree_separation_range=(0.1, 0.7),
    seed=42
)

# Generate graphs (stores in family.graphs)
family.generate_family(n_graphs=30, show_progress=True)

# Access generated graphs and convert to PyG format
print(f"Generated {len(family.graphs)} graphs!")
pyg_graphs = family.to_pyg_graphs(task="community_detection")

Option 2: Via YAML config file

# configs/experiment.yaml
universe_parameters:
  K: 10
  edge_propensity_variance: 0.5
  feature_dim: 16
  center_variance: 1.0
  cluster_variance: 0.3
  seed: 42

family_parameters:
  n_graphs: 100
  min_n_nodes: 25
  max_n_nodes: 200
  min_communities: 3
  max_communities: 7
  homophily_range: [0.1, 0.9]
  avg_degree_range: [2.0, 8.0]
  degree_distribution: "power_law"
  power_law_exponent_range: [2.0, 3.0]
  degree_separation_range: [0.4, 0.8]
  seed: 42

task: "community_detection"

# Use config-driven workflow
import yaml
from graph_universe import GraphUniverseDataset

with open("configs/experiment.yaml") as f:
    config = yaml.safe_load(f)

dataset = GraphUniverseDataset(root="./data", parameters=config)
print(f"Generated dataset with {len(dataset)} graphs!")

Option 3: Interactive Demo

Try GraphUniverse in your browser with real-time parameter tuning and direct dataset download: https://graphuniverse.streamlit.app/

Reproduce Validation Experiment

GraphUniverse includes comprehensive metrics to validate property realization and quantify learnable community signals

Reproduce Validation Analysis in Paper Automatically

python validate_parameter_sensitivity.py --n-random-samples 100 --n-graphs 30

Or manually inspect a generated family

# Validate standard graph properties
family_properties = family.analyze_graph_family_properties()
for property_name in ['node_counts', 'avg_degrees', 'homophily_levels']:
    values = family_properties[property_name]
    print(f"{property_name}: mean={np.mean(values):.3f}")

# Analyze within-graph community signals (fits Random Forest per graph)
family_signals = family.analyze_graph_family_signals()
for signal in ['structure_signal', 'feature_signal', 'degree_signal']:
    values = family_signals[signal]
    print(f"{signal}: mean={np.mean(values):.3f}")

# Measure between-graph consistency
family_consistency = family.analyze_graph_family_consistency()
for metric in ['structure_consistency', 'feature_consistency', 'degree_consistency']:
    value = family_consistency[metric]
    print(f"{metric}: {value:.3f}")