π§ Advanced Pattern Evaluation (Node-Based Schema)
This schema-based example walks through how a PET system interprets a recurring set of patterns using the same initial example and the formal architecture:
- ON β Object Node
- PN β Pattern Node
- UP β Understanding Pattern
- SPN β Sub-Pattern Node
- CUG β Current Understanding Graph
- HUG β Historical Understanding Graph
π Example Scenario: βDog Licks Boy in Parkβ
Day 1
- Input: Sight engine reports a composite signal of
Dog licks boy in park
. - PN1 is generated and linked to 4 SPNs:
Dog
,Lick
,Boy
,Park
. - SPNs are converted (or linked) to ONs.
- These ONs are bound by a relationship edge set (EN).
- A new UP1 is formed in the CUG:
- PN1 + {ON_Dog, ON_Lick, ON_Boy, ON_Park}
- Trust initialized (100%)
- Evaluation:
UP1 => Exists
Day 2
- PN2 with
Dog licks boy on street
is input. - SPNs extracted:
Dog
,Lick
,Boy
,Street
- All match existing ONs except
Street
(new ON added) - PN2 evaluated as similar to UP1.
- Reinforces UP1:
- Trust level increases
- Edge added for location variance
- No new UP is created. PN2 is archived in HUG, referenced by UP1.
Day 3
- PN3 with
Dog licks boy in park
again. - Full match with UP1 and ON set.
- Reinforcement increases further.
- No contradiction or ambiguity.
- Archived in HUG.
Day 4 β Contradiction
- PN4:
Dog bites boy in car
- ONs: Dog, Bite (new), Boy, Car (new)
- Mapped to new UP2 in CUG:
- Represents contradiction to UP1
- Recursion Engine activates:
- Evaluates impact on UP1:
- Location (car vs. park)
- Action (bite vs. lick)
- Partial overlap with ONs (Dog, Boy)
- Adjusts trust in UP1
- Lick re-evaluated across UPs and ONs
- Evaluates impact on UP1:
π§ Recap of Node Dynamics
Element | Description |
---|---|
PN | Individual pattern instance |
ON | Canonical object/action node |
UP | Generalized pattern of understanding |
SPN | Extracted sub-nodes from sensor input |
EN | Edges that bind ONs in a relational context |
CUG | Graph of current trusted UPs |
HUG | Archive of past PN evaluations |
π§ Concept Refinement from Contradictions
βLickβ is now connected across:
Dog licks boy
(affection)Tiger licks cub
(grooming)Dog licks wound
(healing)Team licks opponent
(domination)
From these:
Lick
ON forms higher abstraction:- β Affection (exists)
- β Grooming (exists)
- β Domination (context-dependent)
π§ Summary
- Every PN contributes to memory (HUG) and may refine or challenge UPs (CUG).
- Contradictions lower trust, generate new UPs, or reclassify ON meanings.
- Continuity is recursively evaluated from stable UPs reinforced over time.
- New understanding often emerges from contradictions, not repetition.
Last updated: 2025-06-12