Categorical foundations for AI meta-prompting with proven correctness and measurable quality improvements (100% on Game of 24)
-
Updated
Dec 29, 2025 - Python
Categorical foundations for AI meta-prompting with proven correctness and measurable quality improvements (100% on Game of 24)
Quantum NLP sentence classification (UMD FIRE) — pregroup-grammar quantum circuits; enhanced SPSA/ADAM/genetic optimizers beat a default-gain SPSA baseline by ~30 points of train/test accuracy.
Executable categorical models of linguistic case: synthetic statistical studies, DisCoPy diagrams, Bayesian and distributional utilities, bounded MCP tools, and a reproducible manuscript with explicit mathematical and empirical limits.
To associate your repository with the discopy topic, visit your repo's landing page and select "manage topics."