Mercurial > pylearn
comparison doc/v2_planning/sampler.txt @ 1027:a1b6ccd5b6dc
few comments added
author | Yoshua Bengio <bengioy@iro.umontreal.ca> |
---|---|
date | Fri, 03 Sep 2010 21:57:22 -0400 |
parents | 790376d986a3 |
children | 875d53754bd0 |
comparison
equal
deleted
inserted
replaced
1026:38f799f8b6cd | 1027:a1b6ccd5b6dc |
---|---|
1 | |
2 Inference / Sampling committee: JB, GD, AC | |
3 | |
1 OVERVIEW | 4 OVERVIEW |
2 ======== | 5 ======== |
3 | 6 |
4 Before we start defining what a sampler is and how it should be defined in | 7 Before we start defining what a sampler is and how it should be defined in |
5 pylearn, we should first know what we're up against. | 8 pylearn, we should first know what we're up against. |
34 * Metropolis Hastings | 37 * Metropolis Hastings |
35 * Slice Sampling | 38 * Slice Sampling |
36 * Annealing | 39 * Annealing |
37 * Parallel Tempering, Tempered Transitions, Simulated Tempering | 40 * Parallel Tempering, Tempered Transitions, Simulated Tempering |
38 * Nested Sampling (?) | 41 * Nested Sampling (?) |
39 * Hamiltonian Monte Carlo | 42 * Hamiltonian Monte Carlo --> or is it Hybrid Monte Carlo? |
43 | |
44 3. USAGE PATTERNS | |
45 ================= | |
46 | |
47 * MCMC methods have a usage pattern that is quite different from the kind of univariate sampling methods | |
48 needed for nice-and-easy parametric families. | |
49 | |
50 |