Research
Primary publications behind Genefold’s methods. Each paper page states what the work contributes, what is reproducible, and where the canonical version lives. Results are published together with their evaluation protocols.
ArrowSpace: Spectral Search for Embeddings and Graph Analysis
Introduces spectral indexing with lambda-tau scores: Rayleigh-quotient smoothness blended with edge-wise dispersion into bounded, comparable spectral scores.
2026 · Empirical studyEpiplexity and Graph Wiring
An empirical study of how graph-construction parameters shape a generic spectral algorithm: every dataset generates information, every manifold draws a unique surface.
2026 · arXivFrom Embedding Geometry to Spectral Search
Introduces the Graph Wiring framework, Spectral Indexing (SPIN), and tau-modulation, with evaluation across benchmark and industrial settings.
Reproducibility policy
Every benchmark claim on this site is tied to a stated protocol. Where a result is cited, the linking page lists dataset, configuration, and metric. Code for the methods is open source: pyarrowspace and arrowspace-rs.