Network hop

As explained in Graph Neural Networks in Action

A graph hop is one traversal from a node to an adjacent node along an edge.

A network hop is one traversal from a graph node to an adjacent node along an edge. In graph neural networks, hops determine how far neighborhood information can influence a node embedding. section 2.4.4

Message passing connects hops to GNN implementation:

• Node features, edge features, or both form messages passed across the graph. section 2.4.4
• Each message-passing layer collects information from nodes or edges one hop away, then aggregates, transforms, and updates representations. section 2.4.4
• Three hidden message-passing layers give a node representation access to information up to three hops away, while neighborhood coverage can grow exponentially. section 2.4.4

Co-purchasing graph

For books connected by co-purchasing edges, a one-hop layer uses immediately connected books; three layers can incorporate information from books three hops away, including node or edge attributes. section 2.4.4

Figure 2.15 Elements of our message-passing layer. Each message-passing layer consists of an aggregation, a transformation, and an update step.

Use the message-passing layer diagram to trace aggregation, transformation, and update across one hop. section 2.4.4

How Graph Neural Networks in Action covers it

How deep it goes

Explains hops through message-passing layers and notes attention can determine depth, without detailing attention mechanisms.

Examples

Uses a book co-purchasing graph and contrasts one-hop access with three-layer, three-hop neighborhood access.

What it assumes

Assumes basic graph nodes, edges, and neighborhoods while introducing hops through message passing.

What you can do after

Evaluate how GNN layer count controls the graph context available to each node representation.

How it connects

Connects network hops to message passing, graph neural networks, node embeddings, graph neighborhoods, and attention.

liveBookmanning.com

Read Graph Neural Networks in Action on liveBook Get the book at manning.com

Also explained in

• 

  liveBookmanning.com
  Acing the CCNA Exam: Volume 1 Fundamentals and ProtocolsA hop is one journey of a message from one network node to the next node along its path to a destination.

Compare how 2 books explain network hop →

Related concepts

• data link layerproduced by · The data link layer is the network-model layer that provides frame delivery and link addressing between nodes connected by the same physical or logical link.
• direct model embeddingaddressed by · A model-serving design in which model execution code is integrated with application logic and runs in the same process.
• neural network layerequivalent to · A neural network layer is a computational stage or component that transforms input representations into outputs within a neural network.
• message passingconfigured by · A technique in which computational entities exchange messages to communicate or propagate information, often coordinating work or updating state.

This explainer was generated by Manning's knowledge graph from the text of Graph Neural Networks in Action. The section links point to the parts of the book it draws on.