liveBook · Manning

Graph convolutional network

Also called GCN

A graph convolutional network is a graph neural network that computes representations by transforming and aggregating information from graph neighborhoods.

How 6 Manning books explain it

Graph Neural Networks in Action

A convolutional graph neural network is a class of graph neural network architectures that use convolution-like neighborhood aggregation.

How it is used
Applies graph-aware node embeddings to product-category prediction, bundle discovery, and comparisons with feature-only baselines.
How deep it goes
Builds and trains a two-layer PyG model, then examines normalized aggregation and spectral convolution without pursuing a rigorous derivation.
Examples
Uses Amazon co-purchases for category prediction and bundles, PinSage for recommendations, and a GCN baseline against tabular models.
What you can do after
Build it from scratch with PyG’s `GCNConv` layers, train it for node classification, and inspect its embeddings.

Read how Graph Neural Networks in Action explains graph convolutional network →

Optimization Algorithms: AI techniques for design, planning, and control problems

A graph convolutional network is a neural network that aggregates information from graph neighborhoods to compute node, edge, or graph representations and predictions.

How it is used
Uses GCNs to generate graph embeddings for prediction and combinatorial optimization.
How deep it goes
Explains neighborhood-based embedding and demonstrates a two-layer GCN; uses a pretrained model rather than training the TSP model from scratch.
Examples
Uses the Cora node-embedding dataset and a traveling salesperson problem pipeline with pretrained-model experimentation.
What you can do after
Apply GCN layers via PyTorch Geometric and experiment with a pretrained model on traveling salesperson problem instances.

Read how Optimization Algorithms: AI techniques for design, planning, and control problems explains graph convolutional network →

Knowledge Graphs and LLMs in Action

A graph convolutional network is a graph-neural-network model that updates node representations by aggregating information from graph neighborhoods.

How it is used
Deploys GCNs as encoders for financial node classification and recommendation link prediction.
How deep it goes
Explains the encoder–decoder workflow and graph-type operators, but does not derive the convolution mathematically.
Examples
Uses anti-money-laundering classification and user–movie recommendation as concrete GCN applications.
What you can do after
Apply it via PyTorch Geometric and evaluate predictions with precision, recall, F1-score, and confusion matrices.

Read how Knowledge Graphs and LLMs in Action explains graph convolutional network →

Fight Fraud with Machine Learning

A graph convolutional network is a neural network that learns node representations by aggregating information from neighboring nodes in graph data.

How it is used
Detects illicit Bitcoin transactions by incorporating payment-flow neighborhoods alongside transaction features.
How deep it goes
Works through a two-layer architecture, averaging aggregation, forward inputs, baseline visualization, training, and test evaluation.
Examples
Uses a social-network gender classifier and Elliptic Bitcoin transactions, including untrained t-SNE predictions and learning curves.
What you can do after
Build it from scratch as a two-layer PyTorch GCN and evaluate illicit-transaction predictions with accuracy and F1.

Read how Fight Fraud with Machine Learning explains graph convolutional network →

Build AI Drug Discovery Pipelines

A graph convolutional network is a graph neural network that updates node representations by aggregating information from neighboring nodes with graph convolution operations.

How it is used
Uses shared neighborhood aggregation to encode molecular and protein graphs for drug-target affinity prediction.
How deep it goes
Explains tensor representation, layer updates, pooling, architecture dimensions, regularization, and evaluation, but defers detailed derivations.
Examples
Demonstrates three-layer molecular and protein branches, PyTorch Geometric pooling, and GCN, GAT, and GIN alternatives.
What you can do after
Builds a graph encoder that produces fixed-size molecule or protein embeddings for affinity prediction.

Read how Build AI Drug Discovery Pipelines explains graph convolutional network →

Machine Learning Algorithms in Depth

A graph convolutional network is a graph neural network that computes node representations by aggregating information from graph neighborhoods.

How it is used
Classifies nodes in relational datasets by combining node features with information propagated through graph neighborhoods.
How deep it goes
Defines graph notation, feature matrices, and a normalized forward equation, then applies the model through an implementation.
Examples
Uses the CORA citation network, binary document-word features, the normalized forward equation, and a TensorFlow implementation.
What you can do after
Build and apply a GCN classifier for citation graphs using TensorFlow, Keras, and Spektral.

Read how Machine Learning Algorithms in Depth explains graph convolutional network →

What to know first

Related concepts

Questions about graph convolutional network

What is graph convolutional network?

A graph convolutional network is a graph neural network that computes representations by transforming and aggregating information from graph neighborhoods.

Which Manning books explain graph convolutional network?

Graph Neural Networks in Action, Optimization Algorithms: AI techniques for design, planning, and control problems, Knowledge Graphs and LLMs in Action, Fight Fraud with Machine Learning, Build AI Drug Discovery Pipelines and Machine Learning Algorithms in Depth.

What should I know before learning graph convolutional network?

Manning books that teach graph convolutional network usually introduce convolutional neural network, graph neural network, dropout and machine learning first.

What is the difference between graph convolutional network and GraphSAGE?

A graph convolutional network is a graph neural network that computes representations by transforming and aggregating information from graph neighborhoods. GraphSAGE is an inductive graph-neural-network method that learns node representations by sampling and aggregating information from node neighborhoods.

What is graph convolutional network a kind of?

Graph convolutional network is a kind of graph neural network. A graph neural network is a neural-network model that learns representations or predictions from graph-structured data by combining information from connected entities.

What is graph convolutional network also called?

Graph convolutional network is also called GCN.

These guides are generated from the text of Manning books by Manning's knowledge graph and link to the sections they draw on. The full books are available on liveBook and at manning.com.