◎ OS PUB Apache 2.0 ← All specifications

P145 — AIEP — Continuous Evidence and Background Learning Protocol

Publication Date: 2026-04-12 Status: Open Source Prior Art Disclosure Licence: Apache License 2.0 Author/Organisation: Phatfella Ltd Schema: AIEP_OS_SPEC_TEMPLATE v1.0.1 — https://aiep.dev/schemas/aiep-os-spec-template/v1.0.1

Field of the Invention

[0001] The disclosure relates to continuous evidence ingestion and knowledge state update systems within computing environments.

[0002] More particularly, the disclosure concerns a continuous evidence and background learning protocol for use within an Architected Instruction and Evidence Protocol (AIEP) system, enabling an AI system to ingest new evidence artefacts continuously, update its knowledge state deterministically, and maintain a cryptographically auditable record of all knowledge state changes without interfering with foreground reasoning operations.

Framework Context

[0003] This invention operates within an Architected Instruction and Evidence Protocol (AIEP) environment as defined in United Kingdom patent application number GB2519711.2, filed 20 November 2025, the entire contents of which are incorporated herein by reference.

[0004] The present invention extends evidence freshness and temporal relevance mechanisms defined in P147, the persistent world model protocol defined in P137, and evidence change notification mechanisms defined in P180 while remaining independently implementable as described herein.

Background

[0005] AI systems operating in dynamic environments must incorporate new evidence continuously without disrupting active reasoning. Uncontrolled background learning introduces hidden state modifications that break deterministic reproducibility, corrupt in-progress reasoning chains, and prevent audit of knowledge state at query time.

[0006] There exists a need for a background evidence ingestion mechanism that isolates background updates from foreground reasoning at the evidence substrate level, applies updates in deterministic order, and records all knowledge state transitions with cryptographic binding.

Summary of the Disclosure

[0007] A computer-implemented continuous evidence and background learning protocol comprises: maintaining a BackgroundIngestionQueue receiving evidence artefacts from configured sources; processing BackgroundIngestionQueue entries in deterministic arrival order without preempting foreground reasoning sessions; for each ingested artefact, evaluating admissibility using the standard AIEP admissibility pipeline; on admissibility confirmation, computing a BackgroundLearningHash as H(prior_knowledge_state_hash || artefact_hash || update_rule_id || ingestion_timestamp); appending a BackgroundLearningRecord to the knowledge state audit log; making the updated knowledge state available to subsequent foreground reasoning sessions after the BackgroundLearningRecord is durably appended; and isolating in-progress foreground reasoning sessions from background state updates by binding them to the knowledge state hash at session initialisation.

[0008] Background ingestion is rate-governed to prevent resource contention with foreground reasoning. Inadmissible artefacts are quarantined, not discarded.

[0009] The technical effect is modification of computing system behaviour by enforcing isolated deterministic background knowledge updates with cryptographic audit records, enabling continuously updated AI knowledge states without foreground reasoning interference.

Claims

[0010] A computer-implemented method for continuous background evidence learning comprising: maintaining a BackgroundIngestionQueue; processing entries in deterministic order without foreground preemption; evaluating admissibility for each artefact; computing BackgroundLearningHashes; appending BackgroundLearningRecords; and isolating foreground sessions to their initialisation-time knowledge state hash.

[0011] The method of claim 1 wherein inadmissible artefacts are quarantined and not discarded, preserving audit coverage of all received evidence.

[0012] The method of claim 1 wherein foreground reasoning session isolation is enforced by binding the session to the knowledge state hash at session initialisation time.

[0013] A system for continuous background evidence learning comprising one or more processors and a non-transitory computer-readable medium storing instructions to execute the method of any preceding claim.