System Intelligence Dashboard

Operational metrics showing how Logicon's knowledge base grows and forecasting calibration improves over time

Last updated: 26 Apr 2026, 03:49

At a Glance

Key performance indicators from production systems

Live
Live
1,374
Predictions Generated
480 backfill resolved
Computed
186
Brain Documents
11 domain branches
Computed
11
Active Agents
Specialized domain experts
Live
0.1059
Brier Score
n=480 (lower = better)
Live
0.9492
AUC-ROC
Discrimination ability
Live
7/9
OSINT Sources
724 sweeps completed

Knowledge Base Growth

Documents added per week, by domain branch

Computed

Top Branches

06 forecasting ml
28
08 geopolitical context
22
01 diana
21
02 nato doctrine
20
07 osint sources
18
04 warfighter operations
15

Forecasting Calibration

Continuous improvement through walk-forward validation and isotonic calibration

Live

Brier Score Trajectory

Lower is better. Target: 0.10

AUC-ROC Trajectory

Higher is better. Ascending = improvement

Cumulative Predictions

Total structural predictions generated over time

Domain Calibration Scores

Brier score by conflict theatre (lower = better calibrated)

Live

Prediction Methods

Structural prediction generation by method

Live

Active Parameters

pitf-imr-v0full_snapshot
default-v0-1777033693869isotonic calibration
defence-active-1777033693953ml defence ensemble
defence-logistic-1777033693955ml confidence
34 total versions · 0 drift alerts

Specialized Agent Network

Domain expert agents with dedicated knowledge bases

Illustrative
Design View — Agent telemetry instrumentation planned for Phase 1. Agent definitions and knowledge bases are operational.
01
forecasting ml
28 docs
02
geopolitical context
22 docs
03
diana
21 docs
04
nato doctrine
20 docs
05
osint sources
18 docs
06
warfighter operations
15 docs
07
defence tech industry
14 docs
08
compliance legal
14 docs
09
mss nato
13 docs
10
logicon internal
11 docs
11
aco sop
8 docs

OSINT Source Operational Status

Independent sources with redundancy and quality monitoring

Live

Last sweep: 24 Apr 2026, 10:18 (82.0s)

alphavantage
Operational
alternative_me
Operational
binance_announcements
Operational
coinmarketcal
Operational
fred
Operational
gdelt
Failed
war: fetch failed | oil_energy: fetch failed | crypto_reg: f
halvings
Operational
openexchange
Operational
opensanctions
Failed
throttled: ran within last 120min

Calibration Diagnostics

Standard ML reliability assessment

Live

Reliability Diagram

Perfect calibration = diagonal line. n=480

Insufficient data for reliability diagram

Prediction Confidence Distribution

Histogram of all backfill prediction probabilities

Learning Pipeline Activity

Autonomous retraining cycles and model activations

Live
gate3 first trainingactivated24/04/2026, 15:28
gate3 first trainingactivated24/04/2026, 15:23
gate3 first trainingactivated24/04/2026, 15:22
gate3 first trainingactivated24/04/2026, 15:17
gate3 first trainingactivated23/04/2026, 01:58
gate3 first trainingrejected23/04/2026, 01:58
gate3 first trainingrejected23/04/2026, 01:56
manualerror21/04/2026, 17:22

System Intelligence Index

Composite metric of operational maturity

Computed
78
/ 100
Calibration
58 (35%)
Discrimination
95 (25%)
Knowledge
93 (20%)
Sources
78 (20%)

Computed daily. Formula: Calibration (35%) + Discrimination (25%) + Knowledge (20%) + Sources (20%). Methodology at /methodology.