Top-level navigation. See AGENTS.md for the schema and how the assistant should use this wiki.
- [[bayesian-inference]] — likelihoods, priors, posteriors, evidence.
- [[model-comparison]] — Bayes factors, AIC/BIC, posterior predictive checks.
- [[expectation-propagation]] — factor graphs, message passing,
PyAutoFit's
graphical/EP framework; 2026 audit findings.
-
[[sampler-candidate-curation]] — bounded source-backed shortlist and on-demand public-provenance handoff to Insight; no scheduled scan.
-
[[mcmc-samplers]] — Metropolis-Hastings, ensemble samplers, Eryn.
-
[[hamiltonian-monte-carlo]] — NUTS, BlackJAX, NumPyro, Stan; gradients via JAX.
-
[[nested-sampling]] — MultiNest, PolyChord, dynesty, UltraNest.
-
[[gpu-nested-sampling]] — JAX-accelerated nested sampling (NSS); vmap batching, chunked vmap.
-
[[initialization-chaining]] — which samplers provide vs require starting points; PyAutoFit chained searches.
-
[[sampler-benchmarks]] — benchmark methodology, historical minimal + A100 results, and pointers to Insight's search campaigns and the project-owned catalogue.
-
[[gradient-optimizer-benchmarks]] — JAX MAP/optimization study on the MGE lens likelihood (CPU + A100): diversity × gradients; multi-start Adam scales (GIGA-Lens), SVGD reaches truth, CMA-ES collapses.
- [[simulation-based-inference]] — SBI, ABC, neural ratio estimators.
- [[neural-posterior-estimation]] — nbi, sbi, lampe, LtU-ILI.
- [[likelihood-emulation]] — neural likelihood surrogates.
- [[probabilistic-programming]] — PPL philosophy, conjugate vs MCMC vs NPE.
- [[message-passing]] — graphical models, belief propagation.
- [[deep-learning-astro]] — CNNs, BNNs, transformers in astrophysics.
- [[active-learning]] — accelerated Bayesian inference via surrogates.
- [[neural-emulators]] — emulating cosmology / N-body / lensing.
- [[linear-inversion]] — least-squares with regularisation.
- [[gaussian-linear-models]] — closed-form Bayesian linear inference.
- [[nufft]] — non-uniform fast Fourier transform.
- [[jax-finufft]] — GPU-accelerated NUFFT.
- [[autodiff-implicit-diff]] — automatic / implicit differentiation through solvers.
- [[jax-ecosystem]] — Equinox, Optimistix, BlackJAX, NumPyro, …
- [[aggregator-performance]] — PyAutoFit result loading at catalogue scale: 2026-07 profiling campaign, sqlite verdict, deserialization floor, bugs found by realistic mock data.
- [[pyautofit]] (cross-link to
../lensing/) — model-fitting framework underlying PyAutoLens. - [[scientific-software]] — general-purpose tooling.
- [[forward-modelling]] — full forward-model pipelines.
- [[hierarchical-population-models]] — multi-level Bayesian inference.
- Samplers / codes: [[polychord]], [[multinest]], [[dynesty]], [[ultranest]], [[nautilus]], [[emcee]], [[blackjax]], [[numpyro]], [[stan]], [[pymc]], [[eryn]], [[nbi]], [[lampe]], [[ltu-ili]].
- JAX tooling: [[optimistix]], [[equinox]], [[jaxopt]], [[jaxns]].
- Astro-method codes: [[pyautofit]], [[pylops]], [[finufft]], [[pynufft]].
These pages describe claim support. Canonical citation metadata and key-management rules
live in ../../bibliography/.
- [[sources-bayesian-inference]]
- [[sources-samplers]]
- [[sources-likelihood-free-inference]]
- [[sources-probabilistic-programming]]
- [[sources-linear-algebra]]
- [[sources-nufft]]
- [[sources-scientific-software]]
- [[sources-simulations]]
- Seed (unverified imports): [[seed-deep-learning-methods]],
[[seed-medical-imaging-adjacents]] — filename-inferred stubs from the 2026-05
PDF import, every entry still reading
Canonical BibTeX key: TODO — no unique match found. They live inseed/, notsources/, and nothing in them may be cited until the paper itself is verified and the entry promoted into the matchingsources/page.
- AGENTS.md — schema and usage rules.
- [[log]] — compilation history.