◎ OS PUB Apache 2.0 ← All specifications

P95 — AIEP — Cross-Session Cognitive Pattern Accumulation Substrate

Publication Date: 2026-03-01 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

Framework Context

[0001] This disclosure 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.

[0002] The present disclosure extends deterministic canonicalisation, governance, and execution integrity mechanisms defined in the AIEP environment while remaining independently implementable as described herein.

Field of the Disclosure

[0003] This disclosure relates to persistent governed reasoning substrates for cognitive assistance systems.

[0004] More particularly, the disclosure concerns a deterministic mechanism within an AIEP substrate for accumulating across sessions a structured representation of a user’s characteristic reasoning patterns — comprising their recurring domain orientations, branch formation tendencies, evidence weighting preferences, and convergence path signatures — and applying this accumulated CognitivePatternProfile to weight the admission priority of incoming evidence, without performing behavioural inference or personalisation outside the constitutional governance framework.

Background

[0005] AI reasoning substrates that persist across user sessions accumulate information about how a specific user engages with evidence and forms reasoning branches over time.

[0006] Conventional personalisation systems model user behaviour through statistical pattern matching on interaction history — what content the user engages with, how long they spend on items, what they click. This produces a behavioural preference model, not a cognitive pattern model.

[0007] A cognitive pattern model is structurally distinct from a behavioural preference model. It represents not what topics a user returns to, but how the user reasons — their characteristic approach to evidence weighting, their tendency toward particular branch formation patterns, their typical convergence paths from evidence to conclusion.

[0008] Existing systems do not provide: extraction of a stable CognitivePatternProfile distinct from topic preferences and behavioural signals; version-bound deterministic computation of CognitivePatternProfile fields from canonical ledger entries; application of the CognitivePatternProfile to modulate evidence admission priority without suppressing admission; cryptographic binding of the CognitivePatternProfile to the specific ledger state from which it was derived; or governance-constrained profile application that reorders but never rejects evidence.

Summary of the Disclosure

[0009] A CognitivePatternProfile is maintained per user comprising four canonical fields:

DomainOrientationVector — a weighted representation of the reasoning domains the user most frequently engages with, computed from Reasoning Ledger domain classification fields.
BranchFormationTendencyVector — a representation of the user’s characteristic patterns in forming new branches versus extending existing ones, computed from branch creation and extension event ratios.
EvidenceWeightingSignature — a representation of the provenance classes, source types, and corroboration patterns most heavily weighted historically, computed from Evidence Ledger weight distributions.
ConvergencePathSignature — a representation of the structural characteristics of reasoning paths by which the user most frequently reaches convergent conclusions.

[0010] CognitivePatternProfile fields are computed deterministically from canonical ledger entries using version-bound schema-defined extraction functions.

[0011] A CognitivePatternProfileHash is computed as:

CognitivePatternProfileHash = H(
  CanonicalSerialise(DomainOrientationVector) ||
  CanonicalSerialise(BranchFormationTendencyVector) ||
  CanonicalSerialise(EvidenceWeightingSignature) ||
  CanonicalSerialise(ConvergencePathSignature) ||
  SchemaVersionId
)

[0012] Upon evidence admission, an EvidenceCongruenceScore is computed as a deterministic function of the artefact’s evidence fields and the active CognitivePatternProfile, modulating the artefact’s position in the admission queue. EvidenceCongruenceScore cannot cause an artefact to be rejected. All constitutionally admissible artefacts are admitted regardless of score. Maximum queue advancement is a schema-defined bounded parameter.

[0013] In hardware-enabled deployments, the CognitivePatternProfile is maintained within the governance chip hardware isolation enclave and is not transmitted to external systems without explicit user-authorised export.

Licence

Licensed under the Apache License, Version 2.0 (the “License”); you may not use this file except in compliance with the License. You may obtain a copy of the License at https://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an “AS IS” BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.