1、Introduction to Sampling based inference and MCMC,Ata Kaban School of Computer Science The University of Birmingham,The problem,Up till now we were trying to solve search problems (search for optima of functions, search for NN structures, search for solution to various problems) Today we try to:- Co
2、mpute volumes Averages, expectations, integrals Simulate a sample from a distribution of given shape Some analogies with EA in that we work with samples or populations,The Monte Carlo principle,p(x): a target density defined over a high-dimensional space (e.g. the space of all possible configuration
3、s of a system under study) The idea of Monte Carlo techniques is to draw a set of (iid) samples x1,xN from p in order to approximate p with the empirical distributionUsing these samples we can approximate integrals I(f) (or v large sums) with tractable sums that converge (as the number of samples gr
4、ows) to I(f),Importance sampling,Target density p(x) known up to a constant Task: compute Idea: Introduce an arbitrary proposal density that includes the support of p. Then:Sample from q instead of p Weight the samples according to their importance It also implies that p(x) is approximated byEfficie
5、ncy depends on a good choice of q.,Sequential Monte Carlo,Sequential: Real time processing Dealing with non-stationarity Not having to store the data Goal: estimate the distrib of hidden trajectories We observe yt at each time t We have a model: Initial distribution: Dynamic model: Measurement model
6、:,Can define a proposal distribution:Then the importance weights are:Obs. Simplifying choice for proposal distribution: Then:,fitness,proposed,weighted,re-sampled,proposed,-,weighted,Applications,Computer vision Object tracking demo Blake&Isard Speech & audio enhancement Web statistics estimation Re
7、gression & classification Global maximization of MLPs Freitas et al Bayesian networks Details in Gilks et al book (in the School library) Genetics & molecular biology Robotics, etc.,M Isard & A Blake: CONDENSATION conditional density propagation for visual tracking. J of Computer Vision, 1998,Refere
8、nces & resources,1 M Isard & A Blake: CONDENSATION conditional density propagation for visual tracking. J of Computer Vision, 1998Associated demos & further papers: http:/www.robots.ox.ac.uk/misard/condensation.html 2 C Andrieu, N de Freitas, A Doucet, M Jordan: An Introduction to MCMC for machine l
9、earning. Machine Learning, vol. 50, pp. 5-43, Jan. - Feb. 2003. Nando de Freitas MCMC papers & sw http:/www.cs.ubc.ca/nando/software.html 3 MCMC preprint service http:/www.statslab.cam.ac.uk/mcmc/pages/links.html 4 W.R. Gilks, S. Richardson & D.J. Spiegelhalter: Markov Chain Monte Carlo in Practice.
10、 Chapman & Hall, 1996,The Markov Chain Monte Carlo (MCMC) idea,Design a Markov Chain on finite state spacesuch that when simulating a trajectory of states from it, it will explore the state space spending more time in the most important regions (i.e. where p(x) is large),Stationary distribution of a
11、 MC,Supposing you browse this for infinitely long time, what is the probability to be at page xi. No matter where you started off.=PageRank (Google),Google vs. MCMC,Google is given T and finds p(x) MCMC is given p(x) and finds T But it also needs a proposal (transition) probability distribution to b
12、e specified. Q: Do all MCs have a stationary distribution? A: No.,Conditions for existence of a unique stationary distribution,Irreducibility The transition graph is connected (any state can be reached) Aperiodicity State trajectories drawn from the transition dont get trapped into cycles MCMC sampl
13、ers are irreducible and aperiodic MCs that converge to the target distribution These 2 conditions are not easy to impose directly,Reversibility,Reversibility (also called detailed balance) is a sufficient (but not necessary) condition for p(x) to be the stationary distribution.It is easier to work w
14、ith this condition.,MCMC algorithms,Metropolis-Hastings algorithm Metropolis algorithm Mixtures and blocks Gibbs sampling other Sequential Monte Carlo & Particle Filters,The Metropolis-Hastings and the Metropolis algorithm as a special case,Obs. The target distrib p(x) in only needed up to normalisa
15、tion.,Examples of M-H simulations with q a Gaussian with variance sigma,Variations on M-H: Using mixtures and blocks,Mixtures (eg. of global & local distributions) MC1 with T1 and having p(x) as stationary p MC2 with T2 also having p(x) as stationary p New MCs can be obtained: T1*T2, or a*T1 + (1-a)
16、*T2, which also have p(x) Blocks Split the multivariate state vector into blocks or components, that can be updated separately Tradeoff: small blocks slow exploration of target p large blocks low accept rate,Gibbs sampling,Component-wise proposal q:Where the notation means:Homework: Show that in thi
17、s case, the acceptance probability is =1 see 2, pp.21,Gibbs sampling algorithm,More advanced sampling techniques,Auxiliary variable samplers Hybrid Monte Carlo Uses the gradient of p Tries to avoid random walk behavior, i.e. to speed up convergence Reversible jump MCMC For comparing models of different dimensionalities (in model selection problems) Adaptive MCMC Trying to automate the choice of q,
