COREX · REAL-TIME DASHBOARD

Causal Evaluation Dashboard

Live monitoring of four-axis robustness metrics · COREX score · Real-time classification
COREX SCORE · LIVE
Last Update: --:--:-- UTC
COREX CAUSAL SCORE
0.00
--
Threshold: CAUSAL ≥ 0.80 · SPURIOUS 0.50–0.79 · ARTIFACT < 0.50
0.00
Statistical (S)
0.00
Representation (R)
0.00
Intervention (I)
0.00
Domain (D)
--
Classification
Four-Axis Evaluation Modules · Real-time Performance
Statistical Stability (S)
Cross-subpopulation conditional invariance · KL divergence
0.00
Representation Invariance (R)
Stability under transformations · PCA · Autoencoder · Noise
0.00
Intervention Consistency (I)
Causal effect simulation · Propensity matching
0.00
Domain Robustness (D)
Cross-environment generalization · CV stability
0.00
📊 COREX Scoring Formula
COREX = w₁·S + w₂·R + w₃·I + w₄·D
w₁=0.25 · Statistical
w₂=0.25 · Representation
w₃=0.30 · Intervention
w₄=0.20 · Domain
⚠️ Decision Thresholds
🟢 CAUSAL
≥ 0.80 · All modules stable · Intervention consistent
🟡 SPURIOUS
0.50 – 0.79 · Domain shift OR intervention instability
🔴 ARTIFACT
< 0.50 · Representation invariance fails
🐙 GitHub 🐍 PyPI Package 📌 DOI 📋 Full Results