liveBook · Manning

Pose data

As explained in Graph Neural Networks in Action

Data describing the configuration or positions of body parts over time.

Pose data records sensor readings across body-part nodes and frames. An NRI encoder compresses these temporal patterns and relationships into low-dimensional edge representations, preserving structure for pose prediction. section 6.4.1

A practical encoding pipeline is:

  • Transform node features into edge features, aggregate incoming edges into nodes, then repeat node-to-edge transformation. section 6.4.1

How Graph Neural Networks in Action covers it

How deep it goes
Explains fixed-window preparation, repeated message passing, graph construction, and PyTorch encoder execution.
Examples
Uses multi-sensor frame sequences, fully connected three-node graphs, edge-index tensors, and encoder forward-pass listings.
What it assumes
Assumes familiarity with message passing and Neural Relational Inference components before implementing the encoder.
What you can do after
Apply via plain PyTorch an encoder that prepares temporal inputs and produces edge-type logits.
How it connects
Connects message passing, graph neural networks, Neural Relational Inference, edge embeddings, node embeddings, and latent space.
Cover of Graph Neural Networks in Action
liveBookmanning.com
Read Graph Neural Networks in Action on liveBook Get the book at manning.com

Related concepts

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.