Edges

In the PET schema, edges are the relationships or connections between two or more nodes. These can represent associations such as cause/effect, classification, experience, or sensory correlation.

Unlike traditional edges which are simply directional links (from → to), PET edges may reference an Understanding Node that explains the reason or source behind the connection.

Structure

An edge may include:

• from: ID or label of the source node
• to: ID or label of the target node
• type: (optional) nature of the relationship (e.g., “is-a”, “part-of”, “experienced-at”)
• understanding_id: reference to the Understanding Node that justifies this edge
• weight: (optional) derived from the neural net to indicate strength or trust
• existence: (optional) flag showing whether this edge currently holds in short-term memory

Example

from: Dog  
to: Animal  
type: is-a  
understanding_id: U102  
weight: 0.98  
existence: true

Notes

• Edges can evolve as the neural network adapts to new patterns or reinterpretations.
• Multiple edges between the same pair of nodes may exist, each backed by different understanding nodes.
• The existence flag is often calculated by the neural network and represents the current state of belief.