Samplinglib · research resources
Sampling
Begin with the target, state space and access model. Choose a sampling regime before a named algorithm; use the textbook for foundations and SampleWiki for research results.
Choose a reading route
Log-Concave Sampling
Chewi's source-ordered textbook and official supplement
SampleWiki
Sampling regimes, research problems and new results
Discrete Sampling
Finite-state models, spectral independence and mixing
Markov Chain Monte Carlo
General-state kernels, scalable methods and estimation
Keep the argument inspectable
Specify the experiment
Record the target law, continuous or discrete state, initialization and permitted oracle. A sampler must define an actual measurable random output or probability kernel.
Separate the proof obligations
Construction, non-explosion, invariance, convergence and implementation bias are different results. Discrete sampling is not the numerical discretization of a continuous process.
Compose errors and costs
Match TV, Wasserstein, KL or another discrepancy to the theorem. Track conditional errors and costs under the actual input law; TV proximity alone does not control an unbounded expected query count.
Optimization primitives can support a sampling proof but cannot replace kernel semantics, integrability, domain assumptions, a mixing estimate or a cost theorem.
Research projects
Sampling regimes remain the primary SampleWiki entrance; these paper readers are a further drill-down, not a competing library.
Where optimization foundations can help
Convexity, smoothness, proximal energy and scalar contraction estimates can be shared when their definitions and hypotheses actually agree. A Markov kernel, stationarity, an invariant law and an expected-cost bound still need separate proofs.
Compare the upstream primitives and their strict boundaries · Existing research-to-textbook placements