Workflows

Version 1.1 by Robert Schaub on 2025/12/11 16:36

Workflows

This chapter defines the core workflows used across the FactHarbor system.

Each workflow describes:

  • Purpose
  • Participants
  • Steps
  • Automation vs. manual work
  • (Wherever applicable) linear ASCII flow: a → b → c → d

Workflows included:

  1. Claim Workflow
    2. Scenario Workflow
    3. Evidence Workflow
    4. Verdict Workflow
    5. Re-evaluation Workflow
    6. Federation Synchronization Workflow
    7. User Role & Review Workflow
    8. AKEL Workflow
    9. Global Trigger Flow
    10. Entity Lifecycle Notes

Claim Workflow

Purpose:  
Transform raw text or input material into a normalized, classified, deduplicated, and versioned claim ready for scenario evaluation.

Participants:  

  • Contributor  
  • AKEL  
  • Reviewer  

Steps

      1. Ingestion  
  • User submits text, URL, transcript, or multi-claim content  
  • AKEL extracts one or multiple claims  

2. Normalization  

  • Standardizes wording  
  • Reduces ambiguity  
  • Flags implicit assumptions  

3. Classification (AKEL draft → Reviewer confirm)  

  • ClaimType  
  • Domain  
  • Evaluability  
  • SafetyCategory  

4. Duplicate & Similarity Detection  

  • Embeddings created  
  • Similar claims found  
  • Reviewer merges, splits, or confirms uniqueness  

5. Version Creation  

  • New ClaimVersion stored  
  • Every edit creates a new immutable version  

6. Cluster Assignment  

  • AKEL proposes cluster membership  
  • Reviewer confirms  

7. Scenario Linking (optional)  

  • Existing scenarios connected  
  • AKEL may propose new drafts  

8. Publication  

  • Claim becomes active and visible  

Flow:  
Ingest → Normalize → Classify → Deduplicate → Cluster → Version → Publish


Scenario Workflow

Purpose:  
Define the specific analytic contexts needed to evaluate each claim.

Participants:  

  • Contributor  
  • Reviewer  
  • Domain Expert  
  • AKEL  

Steps

      1. Scenario Proposal  
  • Drafted by contributor or generated by AKEL  

2. Completion of Required Fields  
Must include:  

  • Definitions  
  • Assumptions  
  • ContextBoundary  
  • EvaluationMethod  
  • SafetyClass  
  • VersionMetadata  

3. Safety Interception (AKEL)  
Flags:  

  • non-falsifiable structures  
  • pseudoscientific assumptions  
  • unsafe contexts  

4. Redundancy & Conflict Check  

  • Similar scenarios merged  
  • Contradictions flagged  

5. Reviewer Validation  
Ensures clarity, neutrality, and methodological validity.

6. Expert Approval (mandatory for high-risk domains)  

7. Version Storage  

  • Each revision = new ScenarioVersion  

Flow:  
Draft → Validate → Safety Check → Review → Expert Approval → Version → Activate


Evidence Workflow

Purpose:  
Structure, classify, validate, version, and link evidence to scenarios.

Participants:  

  • Contributor  
  • Reviewer  
  • Domain Expert (when needed)  
  • AKEL  

Steps

      1. Evidence Submission  
  • File, dataset, URL, or extracted text  

2. Metadata Extraction (AKEL)  

  • EvidenceType  
  • Category  
  • Provenance  
  • Study design  
  • ExtractionMethod  
  • ReliabilityHints  

3. Relevance Check  
Reviewer verifies which scenarios the evidence applies to.

4. Reliability Assessment  

  • AKEL proposes reliability  
  • Reviewer confirms  
  • Expert review for complex papers  

5. ScenarioEvidenceLink Creation  
Each link stores:  

  • relevance score  
  • justification  
  • evidence version  

6. Versioning  

  • Any update = new EvidenceVersion  

Flow:  
Submit → Extract Metadata → Evaluate Relevance → Score Reliability → Link → Version


Verdict Workflow

Purpose:  
Generate likelihood estimates per scenario based on evidence and scenario structure.

Participants:  

  • AKEL (drafts)  
  • Reviewer  
  • Domain Expert  

Steps

      1. Evidence Aggregation  
        Collect relevant evidence versions.

2. Draft Verdict Generation (AKEL)  
Outputs:  

  • likelihood range  
  • uncertainty factors  
  • conflict detection  
  • sensitivity analysis  

3. Reasoning Draft  
Structured explanation chain generated by AKEL.

4. Reviewer Validation  
Ensures logic, evidence fit, no hallucinations.

5. Expert Review  
Required for:  

  • medicine  
  • psychology  
  • engineering  
  • political misinformation  
  • controversial or risky domains  

6. Verdict Storage  

  • Every update creates a new VerdictVersion  

Flow:  
Aggregate → Draft Verdict → Draft Explanation → Review → Expert Approval → Version


Re-evaluation Workflow

Purpose:  
Keep verdicts current when evidence or scenarios change.

Trigger Types

  • Evidence updated, disputed, retracted  
  • Scenario assumptions changed  
  • Claim reclassification  
  • AKEL contradiction detection  
  • Federation sync  

Steps

      1. Trigger Detection  
        Re-evaluation engine receives event.

2. Impact Analysis  
Find affected:  

  • scenarios  
  • evidence links  
  • verdicts  

3. AKEL Draft Re-calculation  
New:  

  • likelihood  
  • reasoning  
  • uncertainty  

4. Reviewer Validation  
5. Expert Review (high-risk)  
6. Version Storage  

Flow:  
Trigger → Analyze → Recompute → Review → Expert → Version


Federation Synchronization Workflow

Purpose:  
Exchange structured data between nodes.

Steps

  

  1. Detect version changes  
  2. Build bundle (diff + Merkle tree + signatures)  
  3. Push to peers  
  4. Validate lineage + hashes  
  5. Resolve conflicts (merge or branch)  
  6. Optional re-evaluation  

Flow:  
Change → Bundle → Push → Validate → Merge/Fork → Update


User Role & Review Workflow

Purpose:  
Ensure correctness, neutrality, safety, and resistance to manipulation.

Steps

  

      1. Submission  
        Claim / scenario / evidence / verdict.

2. Auto-check (AKEL)  
Flags unsafe content, contradictions, format issues.

3. Reviewer Validation

4. Expert Validation  
Required for sensitive domains.

5. Moderator Oversight  
Triggered by suspicious behavior.

Flow:  
Submit → Auto-check → Review → Expert → Moderator (if needed)


AKEL Workflow

Purpose:  
Support extraction, drafting, structuring, and contradiction detection.

Stages

A — Input Understanding:  
Extraction, normalization, classification.

B — Scenario Drafting:  
Definitions, boundaries, assumptions.

C — Evidence Processing:  
Retrieval, summarization, ranking.

D — Verdict Drafting:  
Likelihood, explanations, uncertainties.

E — Safety & Integrity:  
Contradictions, hallucination detection.

F — Human Approval:  
Reviewer and/or expert.

Flow:  
Input → Drafts → Integrity → Human Approval


Global Trigger Flow (Cascade)

Trigger Sources:  

  • Claim change  
  • Scenario change  
  • Evidence change  
  • Verdict contradiction  
  • Federation update  
  • AKEL model improvements  

Cascade Flow:  
Trigger → Dependency Graph → Re-evaluation → Updated Verdicts