This section provides detailed explanations of the mathematical and theoretical concepts underlying AGISystem2. Each page is designed to be accessible to software engineers who may not have a formal background in abstract algebra or information theory.
Why XOR binding supports cancellation (and why that matters for UNBIND).
Set-based similarity used by SPHDC for robust comparisons.
Sketching for fast, approximate Jaccard estimation.
Deterministic atom vectors from names for reproducibility.
The algebraic structure with two elements (0 and 1) that makes XOR operations cancellable. Foundation for reversible binding in HDC.
Set-based similarity measure: intersection over union. Used by SPHDC for comparing sparse polynomial representations.
Deterministic vector generation from concept names. Enables reproducibility without explicit storage.
Locality-sensitive hashing for efficient Jaccard estimation. Enables SPHDC sparsification while preserving similarity.