6533b85ffe1ef96bd12c25f7
RESEARCH PRODUCT
Anti-tempered Layered Adaptive Importance Sampling
David LuengoLuca MartinoVictor Elvirasubject
Mathematical optimizationRejection samplingSlice sampling020206 networking & telecommunicationsMarkov chain Monte Carlo02 engineering and technology01 natural sciencesStatistics::ComputationHybrid Monte Carlo010104 statistics & probabilitysymbols.namesakeMetropolis–Hastings algorithm[INFO.INFO-TS]Computer Science [cs]/Signal and Image Processing0202 electrical engineering electronic engineering information engineeringsymbolsParallel tempering0101 mathematicsParticle filter[SPI.SIGNAL]Engineering Sciences [physics]/Signal and Image processingImportance samplingComputingMilieux_MISCELLANEOUSMathematicsdescription
Monte Carlo (MC) methods are widely used for Bayesian inference in signal processing, machine learning and statistics. In this work, we introduce an adaptive importance sampler which mixes together the benefits of the Importance Sampling (IS) and Markov Chain Monte Carlo (MCMC) approaches. Different parallel MCMC chains provide the location parameters of the proposal probability density functions (pdfs) used in an IS method. The MCMC algorithms consider a tempered version of the posterior distribution as invariant density. We also provide an exhaustive theoretical support explaining why, in the presented technique, even an anti-tempering strategy (reducing the scaling of the posterior) can be beneficial. Numerical results confirm the advantages of the proposed scheme.
| year | journal | country | edition | language |
|---|---|---|---|---|
| 2017-08-24 |