◎ OS PUB Apache 2.0 ← All specifications

P136 — AIEP — Hypothesis Fork and Idea Competition 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 deterministic reasoning branch management systems within computing environments.

[0002] More particularly, the disclosure concerns a hypothesis fork and idea competition protocol for use within an Architected Instruction and Evidence Protocol (AIEP) system, enabling competing hypotheses to be evaluated in parallel under deterministic evidence-bound scoring without heuristic elimination, semantic preference, or undisclosed prioritisation.

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 branch pruning mechanisms defined in P47, dissent engine mechanisms defined in P126, and outlier generation mechanisms defined in P49 while remaining independently implementable as described herein.

Background

[0005] Reasoning systems encounter problems admitting multiple competing explanatory hypotheses. Selecting among competing hypotheses requires a transparent, reproducible evaluation mechanism that does not introduce hidden preferences.

[0006] Existing systems select among hypotheses using confidence scores, learned priors, attention weighting, or opaque ranker models. These approaches produce selections that cannot be reproduced from evidence alone and cannot be audited for bias or preference injection.

[0007] There exists a need for a deterministic hypothesis competition mechanism that scores each hypothesis against a shared evidence set using explicit criteria, produces cryptographically bound competition records, and selects the winning hypothesis by deterministic rule without heuristic tie-breaking.

Summary of the Disclosure

[0008] A computer-implemented hypothesis fork and competition method comprises: receiving a HypothesisSet containing two or more candidate hypotheses each bound to an evidence substrate; scoring each hypothesis using a version-bound HypothesisScoringRuleset retrieved from the governance registry; computing a HypothesisCompetitionHash as H(hypothesis_1_hash || score_1 || hypothesis_2_hash || score_2 || … || ruleset_version_id); selecting the winning hypothesis by the highest deterministic score; in the event of a tie, applying the registered tie-breaking rule without heuristic override; generating a HypothesisCompetitionRecord binding all participating hypothesis hashes, scores, ruleset version, and the winning hypothesis identifier; and returning the winning hypothesis with its HypothesisCompetitionRecord as an admissibility-bound artefact.

[0009] Hypotheses eliminated during competition are retained in the HypothesisCompetitionRecord as auditable evidence of alternatives considered.

[0010] The technical effect is modification of computing system behaviour by enforcing transparent evidence-bound hypothesis selection without heuristic preference, enabling independently reproducible reasoning competition outcomes.

Claims

[0011] A computer-implemented method for hypothesis fork and competition comprising: receiving a HypothesisSet; scoring each hypothesis using a version-bound HypothesisScoringRuleset; computing a HypothesisCompetitionHash over all hypotheses, scores, and ruleset version; selecting the winning hypothesis by deterministic rule; and generating a HypothesisCompetitionRecord binding all competition artefacts.

[0012] The method of claim 1 wherein eliminated hypotheses are retained in the HypothesisCompetitionRecord without truncation or summarisation.

[0013] The method of claim 1 wherein tie-breaking is performed using a registered deterministic rule without heuristic override.

[0014] A system for hypothesis fork and competition comprising one or more processors and a non-transitory computer-readable medium storing instructions to execute the method of any preceding claim.