◎ OS PUB Apache 2.0 ← All specifications

P190 — AIEP — Evidence Interoperability Bridge Protocol

Publication Date: 2026-03-27 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 defines a protocol for bidirectional evidence exchange between AIEP evidence corpora and external, non-AIEP evidence repositories — enabling AIEP reasoning chains to consume evidence from external databases (academic repositories, open-data portals, regulatory databases) and enabling external systems to consume AIEP evidence exports, through a structured EvidenceBridge adapter layer that translates between AIEP evidence schemas and external repository schemas while preserving integrity and provenance.

Field of the Disclosure

[0003] This disclosure relates to interoperability protocols for evidence exchange between AIEP-governed evidence corpora and heterogeneous external evidence repositories.

[0004] More particularly, the disclosure concerns: an EvidenceBridge adapter architecture; an inbound translation pipeline converting external records to AIEP EvidenceNodes; an outbound translation pipeline exporting AIEP artefacts to external formats; schema mapping and translation loss detection; bridge integrity attestation; and the relationship between the Bridge Protocol and the normalisation protocol (P10/P17), the provenance protocol (P150), and the federated network protocol (P156).

Background

[0005] AIEP evidence corpora do not exist in isolation. The broader research and regulatory ecosystem produces evidence in heterogeneous schemas: Dublin Core metadata, CrossRef DOI records, FHIR clinical data, SPARQL-queryable linked data triple stores, proprietary licensed database exports. An AIEP node that can bridge to these external sources extends its evidence base without requiring the external sources to adopt AIEP schemas.

[0006] Bridging introduces schema translation risk: fields that exist in AIEP schemas may have no counterpart in the external schema; fields in the external schema may map ambiguously to AIEP fields. Translation loss — information present in the source that cannot be faithfully represented in the target schema — must be detected and recorded, not silently discarded.

Summary of the Disclosure

[0007] EvidenceBridge Architecture: An EvidenceBridge is a configured adapter connecting an AIEP evidence corpus to a specific external repository:

bridge_id — SHA-256 of canonical serialisation of configuration
bridge_name — human-readable identifier
external_repository_type — one of: CROSSREF, PUBMED, ARXIV, DATACITE, SPARQL_ENDPOINT, FHIR_SERVER, CUSTOM:{schema_identifier}
external_schema_version — version of the external schema being bridged
aiep_schema_mapping — a SchemaMapping document (see [0008]) defining field-by-field translations
bridge_direction — INBOUND_ONLY, OUTBOUND_ONLY, BIDIRECTIONAL
last_sync_at — ISO 8601 timestamp of last synchronisation
bridge_node_fingerprint — P46 fingerprint of the AIEP node operating the bridge

[0008] SchemaMapping Document: A SchemaMapping defines, for each external schema field:

external_field — the external schema field name (e.g. Dublin Core dc:title)
aiep_target — the AIEP field it maps to (e.g. citation.title)
transform — optional transformation applied during mapping (e.g. DATE_ISO_8601_PARSE for date fields)
loss_on_absent — NONE (AIEP field is optional and can be null if external field is absent), ADVISORY (AIEP field is expected; absence logged as translation advisory), CRITICAL (AIEP field is required; absence blocks admission)
unmapped_source_fields — a list of external schema fields with no AIEP target; these are preserved verbatim in a bridge_passthrough field on the EvidenceNode and do not contribute to any scored dimensions

[0009] Inbound Translation Pipeline: When an AIEP corpus ingests evidence from an external repository via an EvidenceBridge:

The external record is fetched and deserialized
The SchemaMapping is applied field-by-field to produce a partial EvidenceNode candidate
CRITICAL loss fields are checked; records with missing critical fields are rejected with a BridgeTranslationFailure record
unmapped_source_fields are preserved in bridge_passthrough
The EvidenceNode candidate is passed to the standard inbound pipeline: NORMALISE (P10/P17), CLASSIFY (P160), trust scoring (P124), ProvenanceChain creation (P150) with step type TRANSFER referencing the bridge_id
A BridgeAdmissionRecord is created recording the bridge_id, external record identifier, and translation advisory count

[0010] Outbound Translation Pipeline: When an AIEP artefact is exported to an external repository via a BIDIRECTIONAL bridge:

The AIEP EvidenceNode is retrieved
The SchemaMapping is applied in reverse (AIEP → external field)
Fields in the AIEP schema with no external target are preserved in the external record’s extension namespace where available
An outbound TRANSFER ProvenanceChain step is recorded for the artefact, indicating it was exported to the external repository
The exported record is submitted to the external repository’s ingest API

[0011] Bridge Integrity Attestation: Periodically (default: daily) the bridge node produces a BridgeIntegrityAttestation:

count of records successfully ingested since last attestation
count of BridgeTranslationFailure records
count of translation advisories (ADVISORY loss events)
SHA-256 of the SchemaMapping document (confirming the mapping has not changed)
bridge_node_signature

[0012] Translation Loss Reporting: The BridgeIntegrityAttestation is included in the Evidence Quality Metrics corpus health report (P172), giving operators visibility into the proportion of external evidence admitted with translation advisories.

ASCII Architecture

External Repository (CrossRef, PubMed, SPARQL, FHIR, etc.)
        │  external records
        ▼
EvidenceBridge
  (bridge_id, SchemaMapping, bridge_direction)
        │
        ├── Inbound:
        │   SchemaMapping.apply(external_record) → EvidenceNode candidate
        │   CRITICAL loss? → BridgeTranslationFailure
        │   pass to NORMALISE (P10/P17), CLASSIFY (P160)
        │   ProvenanceChain TRANSFER step (P150)
        │   BridgeAdmissionRecord
        │
        └── Outbound (BIDIRECTIONAL only):
            SchemaMapping.apply_reverse(EvidenceNode) → external record
            ProvenanceChain outbound TRANSFER step
            submit to external repository API
        │
        ▼
BridgeIntegrityAttestation (daily)
  → Quality Metrics (P172)
  → AuditLog (P171)

Operational Detail

[0013] Version Mismatch Handling: External schemas evolve. When an external repository updates its schema version, the external_schema_version in the EvidenceBridge configuration is compared against the version declared in incoming records. Version mismatches trigger a SchemaMismatchAlert and temporarily pause bridging until the SchemaMapping document is updated and validated for the new schema version.

[0014] Deduplication: The inbound pipeline checks the DEID of each candidate EvidenceNode against the existing corpus (P162) before admission. Where an externally-sourced artefact has previously been admitted via a different path (e.g. direct ingestion), the bridge admission is recorded as a TRANSFER step on the existing artefact’s ProvenanceChain rather than creating a duplicate artefact.

[0015] Custom Bridge Extensions: The CUSTOM:{schema_identifier} bridge type allows AIEP operators to define entirely novel SchemaMapping documents for proprietary or specialised external schemas without requiring changes to the bridge core. Custom bridges require explicit governance approval (P89) before activation.

Claims-Exclusion Notice

This specification is published as open-source prior art. No patent claims are asserted by the author in respect of the mechanisms described. Any third party seeking to patent mechanisms substantially equivalent to those described herein is placed on notice of this prior art disclosure.