Publications
Research outputs adapted for public consumption.
Substack
The Knowledge Graph That Taught Itself to Converge
10,467 entities, 289 identity bridges, cross-domain convergence 0.5095. How GOURMET's knowledge graph grew 26x and what it revealed.
ReadFrom Poor to Excellent: How GOURMET Learned to Calibrate Its Predictions
The story of how a prediction system went from Brier 0.270 (Poor) to 0.041 (Excellent) — and what it learned about calibration along the way.
ReadPapers & Reports
Available for Publication
The Knowledge Graph That Taught Itself to Converge
10,467 entities, 289 identity bridges, cross-domain convergence 0.5095. How GOURMET's knowledge graph grew 26x and what it revealed.
From Poor to Excellent: How GOURMET Learned to Calibrate Its Predictions
The story of how a prediction system went from Brier 0.270 (Poor) to 0.041 (Excellent) — and what it learned about calibration along the way.