Samplinglib
Learn the mathematics.
Explore the proofs.
Sampling, optimisation and geometry — connected by formal mathematics.
Lean Graph
Locate the formal results and their checked structure.
Functor Hypergraph
Compare recurring proof mechanisms and their boundaries.
Libraries
Read the mathematics across seven connected subjects.
Benchmarks
Mathematical tasks, proof evidence and evaluation tools.
Early formalisation. Inspect the evidence for each result. Current progress →
Navigation sketches are illustrative; graph views distinguish formal edges from conceptual overlays.
News
Recent updates first. Scroll within the list for earlier news.
- Added subject guides and a pinned research resource directory, drawing on Optlib and Optimization × AI.
- Made Samplinglib's original code and exposition available under Apache-2.0, with explicit third-party boundaries.
- Started SamplingOptimisationGeometryBench, pairing natural-language and Lean QA.
- Added statement seals, proof digestion and evidence-routed source-fidelity review.
- Added source-grounded research workspaces, proof ledgers, and a downloadable ChatGPT/MCP App.
- Our system paper was accepted to the NeurIPS 2026 Workshop MATH-AI.
- Compiled and source-reviewed the first SPHMC and Proximal BPS proof packets.
- Added Discrete Sampling and MCMC as peer libraries.
- Added Statistical Optimal Transport and the Functor Hypergraph.
- Added source-fidelity checks and theorem denoising.
- Established Auto-Sampling-Theory-In-Sleep.
