Large sample asymptotics of the pseudo-marginal method
The pseudo-marginal algorithm is a variant of the Metropolis--Hastings algorithm which samples asymptotically from a probability distribution when it is only possible to estimate unbiasedly an unnormalized version of its density. Practically, one has to trade-off the computational resources used to...
| Main Authors: | , , , |
|---|---|
| Format: | Journal article |
| Language: | English |
| Published: |
Oxford University Press
2020
|
| _version_ | 1826259796748140544 |
|---|---|
| author | Schmon, S Deligiannidis, G Doucet, A Pitt, M |
| author_facet | Schmon, S Deligiannidis, G Doucet, A Pitt, M |
| author_sort | Schmon, S |
| collection | OXFORD |
| description | The pseudo-marginal algorithm is a variant of the Metropolis--Hastings algorithm which samples asymptotically from a probability distribution when it is only possible to estimate unbiasedly an unnormalized version of its density. Practically, one has to trade-off the computational resources used to obtain this estimator against the asymptotic variances of the ergodic averages obtained by the pseudo-marginal algorithm. Recent works optimizing this trade-off rely on some strong assumptions which can cast doubts over their practical relevance. In particular, they all assume that the distribution of the difference between the log-density and its estimate is independent of the parameter value at which it is evaluated. Under regularity conditions we show here that, as the number of data points tends to infinity, a space-rescaled version of the pseudo-marginal chain converges weakly towards another pseudo-marginal chain for which this assumption indeed holds. A study of this limiting chain allows us to provide parameter dimension-dependent guidelines on how to optimally scale a normal random walk proposal and the number of Monte Carlo samples for the pseudo-marginal method in the large-sample regime. This complements and validates currently available results. |
| first_indexed | 2024-03-06T18:55:27Z |
| format | Journal article |
| id | oxford-uuid:11ad2741-6268-4cfd-b2dd-4874783911f7 |
| institution | University of Oxford |
| language | English |
| last_indexed | 2024-03-06T18:55:27Z |
| publishDate | 2020 |
| publisher | Oxford University Press |
| record_format | dspace |
| spelling | oxford-uuid:11ad2741-6268-4cfd-b2dd-4874783911f72022-03-26T10:03:40ZLarge sample asymptotics of the pseudo-marginal methodJournal articlehttp://purl.org/coar/resource_type/c_dcae04bcuuid:11ad2741-6268-4cfd-b2dd-4874783911f7EnglishSymplectic ElementsOxford University Press2020Schmon, SDeligiannidis, GDoucet, APitt, MThe pseudo-marginal algorithm is a variant of the Metropolis--Hastings algorithm which samples asymptotically from a probability distribution when it is only possible to estimate unbiasedly an unnormalized version of its density. Practically, one has to trade-off the computational resources used to obtain this estimator against the asymptotic variances of the ergodic averages obtained by the pseudo-marginal algorithm. Recent works optimizing this trade-off rely on some strong assumptions which can cast doubts over their practical relevance. In particular, they all assume that the distribution of the difference between the log-density and its estimate is independent of the parameter value at which it is evaluated. Under regularity conditions we show here that, as the number of data points tends to infinity, a space-rescaled version of the pseudo-marginal chain converges weakly towards another pseudo-marginal chain for which this assumption indeed holds. A study of this limiting chain allows us to provide parameter dimension-dependent guidelines on how to optimally scale a normal random walk proposal and the number of Monte Carlo samples for the pseudo-marginal method in the large-sample regime. This complements and validates currently available results. |
| spellingShingle | Schmon, S Deligiannidis, G Doucet, A Pitt, M Large sample asymptotics of the pseudo-marginal method |
| title | Large sample asymptotics of the pseudo-marginal method |
| title_full | Large sample asymptotics of the pseudo-marginal method |
| title_fullStr | Large sample asymptotics of the pseudo-marginal method |
| title_full_unstemmed | Large sample asymptotics of the pseudo-marginal method |
| title_short | Large sample asymptotics of the pseudo-marginal method |
| title_sort | large sample asymptotics of the pseudo marginal method |
| work_keys_str_mv | AT schmons largesampleasymptoticsofthepseudomarginalmethod AT deligiannidisg largesampleasymptoticsofthepseudomarginalmethod AT douceta largesampleasymptoticsofthepseudomarginalmethod AT pittm largesampleasymptoticsofthepseudomarginalmethod |