Expand description
Outcome feedback loop for adaptive entity learning.
Tracks which graph entities contributed to session outcomes (success/partial/failure)
and adjusts their utility_score via exponential moving average.
Phase 1 of Adaptive Entity Learning v2.
Structs§
- Contribution
Stats - Get aggregate contribution stats for an entity.
- Entity
Utility - One entity’s utility score after an outcome was applied to it.
- Feedback
Report - Report from a feedback recording operation.
- Session
Entities - The entities one session is known to have touched.
Enums§
- Outcome
Kind - The result of a task or session outcome.
Constants§
- DEFAULT_
UTILITY - Default utility score for new entities.
Functions§
- get_
utility_ score - Get the current utility score for an entity.
- record_
outcome_ feedback - Record outcome feedback: link retrieved entities to an outcome and update utility scores.
- record_
session_ use - Record that a session touched these entities, without judging the outcome.
- session_
entities - The entities a session touched, as its
contributed_torecords tell it.