Mercurial > pylearn
comparison doc/v2_planning/sampler.txt @ 1018:790376d986a3
initial document for sampling
| author | gdesjardins |
|---|---|
| date | Fri, 03 Sep 2010 15:01:02 -0400 |
| parents | |
| children | a1b6ccd5b6dc |
comparison
equal
deleted
inserted
replaced
| 1004:3977ecd49431 | 1018:790376d986a3 |
|---|---|
| 1 OVERVIEW | |
| 2 ======== | |
| 3 | |
| 4 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. | |
| 6 | |
| 7 The workflow I have in mind is the following: | |
| 8 1. identify the most popular sampling algorithms in the litterature | |
| 9 2. get up to speed with methods we're not familiar with | |
| 10 3. identify common usage patterns, properties of the algorithm, etc. | |
| 11 4. decide on an API / best way to implement them | |
| 12 5. prioritize the algorithms | |
| 13 6. code away | |
| 14 | |
| 15 1.BACKGROUND | |
| 16 ============= | |
| 17 | |
| 18 This section should provide a brief overview of what exists in the litterature. | |
| 19 We should make sure to have a decent understanding of all of these (not everyone | |
| 20 has to be experts though), so that we can *intelligently* design our sampler | |
| 21 interface based on common usage patterns, properties, etc. | |
| 22 | |
| 23 Sampling from basic distributions | |
| 24 * already supported: uniform, normal, binomial, multinomial | |
| 25 * wish list: beta, poisson, others ? | |
| 26 | |
| 27 List of sampling algorithms: | |
| 28 | |
| 29 * inversion sampling | |
| 30 * rejection sampling | |
| 31 * importance sampling | |
| 32 * Markov Chain Monte Carlo | |
| 33 * Gibbs sampling | |
| 34 * Metropolis Hastings | |
| 35 * Slice Sampling | |
| 36 * Annealing | |
| 37 * Parallel Tempering, Tempered Transitions, Simulated Tempering | |
| 38 * Nested Sampling (?) | |
| 39 * Hamiltonian Monte Carlo |