Genko 原稿Back to edition

FIRME: Registry-Anchored Memory Graphs for Evolving, Heterogeneous Investor Knowledge

Financial memory breaks when the same fact arrives in new words. FIRME anchors observations to canonical identities: +7.8 to 12.1 points in correct match rate.

The explainer · 254 words

The claim

A registry keeps investor knowledge attached to one identity across periods.

The problem

Financial knowledge graphs split when the same fact arrives in new words.

Fragmented history stops evidence from accumulating.

How it works

Incoming observations reconcile to canonical relationships.

Each relationship keeps its source, time, and context.

A fixed product-topic ontology links calls to cash-flow signals.

Aliases send recurring observations to the same history.

Deterministic updates summarize reinforcement, inactivity, and reactivation.

Figures

How one observation landsIdentity first, evidence after.

Step 1 of 4

Receive

A new observation arrives from a source.

Reported gains over the baselineRegistry memory against a registry without alias memory.

FIRME improvement

Match rate

7.8 to 12.1 points higher.

Plain knowledge graph against registry memoryWhy identity anchoring matters.

Plain knowledge graph

  • Surface forms split the history.
  • Evidence stops accumulating.
  • The same relationship appears twice.

What changes

Correct match rate rises 7.8 to 12.1 points over the baseline.

A blinded human audit finds 9.0 to 16.5 points of gain.

Reattachment recall improves by 22.5 to 29.0 points.

Limits

The evidence comes from proprietary records and S&P 500 transcripts.

Evaluation uses a fixed judge with fallback review.

The gains are measured against a single registry baseline.

Source · arXiv

More from this edition