Berlin Tech Meetup: The Future of Relational Foundation Models, Systems, and Real-World Applications

Register now:
PyG Guide

Graph ML Concepts Explained

Every concept in graph machine learning, explained with enterprise data examples. From message passing to graph transformers, from over-smoothing to relational deep learning.

120+

Concepts

Yes

Agent-Optimized

Every Page

Enterprise Examples

Training

17 concepts

How to train GNNs effectively. Self-supervised pre-training, contrastive methods, temporal sampling strategies, and avoiding data leakage.

Self-Supervised Learning

Learning representations from graph structure without labeled data.

Contrastive Learning

Learning by pulling similar pairs together and pushing dissimilar pairs apart.

Graph Augmentation

Edge dropping, feature masking, and subgraph sampling for robust training.

Transfer Learning

Applying knowledge from one graph domain to another.

Pre-Training

Large-scale unsupervised training before task-specific fine-tuning.

Fine-Tuning

Adapting a pre-trained model to a specific downstream task with limited labels.

Few-Shot Learning

Making predictions with only a handful of labeled examples.

Zero-Shot Prediction

Predicting on unseen classes or domains without any task-specific training.

In-Context Learning

Conditioning predictions on examples provided at inference time.

Negative Sampling

Generating non-edges for link prediction training. Strategy matters enormously.

Masked Token Prediction

Masking node or edge features and predicting them. BERT-style pre-training for graphs.

Multi-Task Learning

Training on multiple objectives simultaneously for shared representations.

Curriculum Learning

Training on easy examples first, gradually increasing difficulty.

Class Imbalance

Handling skewed label distributions. Critical for fraud where positives are 0.1%.

Temporal Sampling

Sampling neighbors respecting time order to prevent information leakage.

Data Leakage

When future information leaks into training. The silent killer of graph ML projects.

Temporal Split

Train/val/test splits based on time, not random. Essential for production validity.

Applications

16 concepts

Where graph ML creates real-world value. Each concept maps to specific industries and business outcomes.

Skip the theory. Get predictions in seconds.

KumoRFM applies these concepts automatically. You describe the prediction task in one line of PQL. The model handles the rest.