Monte Carlo Markov Chain Algorithms for Sampling Strongly Rayleigh Distributions and Determinantal Point Processes

Nima Anari, Shayan Oveis Gharan, Alireza Rezaei
29th Annual Conference on Learning Theory, PMLR 49:103-115, 2016.

Abstract

Strongly Rayleigh distributions are natural generalizations of product and determinantal probability distributions and satisfy the strongest form of negative dependence properties. We show that the "natural" Monte Carlo Markov Chain (MCMC) algorithm mixes rapidly in the support of a homogeneous strongly Rayleigh distribution. As a byproduct, our proof implies Markov chains can be used to efficiently generate approximate samples of a k-determinantal point process. This answers an open question raised by Deshpande and Rademacher which was studied recently by Kang, Li-Jegelka-Sra, and Rebeschini-Karbasi.

Cite this Paper


BibTeX
@InProceedings{pmlr-v49-anari16, title = {Monte Carlo Markov Chain Algorithms for Sampling Strongly Rayleigh Distributions and Determinantal Point Processes}, author = {Anari, Nima and Oveis Gharan, Shayan and Rezaei, Alireza}, booktitle = {29th Annual Conference on Learning Theory}, pages = {103--115}, year = {2016}, editor = {Feldman, Vitaly and Rakhlin, Alexander and Shamir, Ohad}, volume = {49}, series = {Proceedings of Machine Learning Research}, address = {Columbia University, New York, New York, USA}, month = {23--26 Jun}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v49/anari16.pdf}, url = {https://proceedings.mlr.press/v49/anari16.html}, abstract = {Strongly Rayleigh distributions are natural generalizations of product and determinantal probability distributions and satisfy the strongest form of negative dependence properties. We show that the "natural" Monte Carlo Markov Chain (MCMC) algorithm mixes rapidly in the support of a homogeneous strongly Rayleigh distribution. As a byproduct, our proof implies Markov chains can be used to efficiently generate approximate samples of a k-determinantal point process. This answers an open question raised by Deshpande and Rademacher which was studied recently by Kang, Li-Jegelka-Sra, and Rebeschini-Karbasi. } }
Endnote
%0 Conference Paper %T Monte Carlo Markov Chain Algorithms for Sampling Strongly Rayleigh Distributions and Determinantal Point Processes %A Nima Anari %A Shayan Oveis Gharan %A Alireza Rezaei %B 29th Annual Conference on Learning Theory %C Proceedings of Machine Learning Research %D 2016 %E Vitaly Feldman %E Alexander Rakhlin %E Ohad Shamir %F pmlr-v49-anari16 %I PMLR %P 103--115 %U https://proceedings.mlr.press/v49/anari16.html %V 49 %X Strongly Rayleigh distributions are natural generalizations of product and determinantal probability distributions and satisfy the strongest form of negative dependence properties. We show that the "natural" Monte Carlo Markov Chain (MCMC) algorithm mixes rapidly in the support of a homogeneous strongly Rayleigh distribution. As a byproduct, our proof implies Markov chains can be used to efficiently generate approximate samples of a k-determinantal point process. This answers an open question raised by Deshpande and Rademacher which was studied recently by Kang, Li-Jegelka-Sra, and Rebeschini-Karbasi.
RIS
TY - CPAPER TI - Monte Carlo Markov Chain Algorithms for Sampling Strongly Rayleigh Distributions and Determinantal Point Processes AU - Nima Anari AU - Shayan Oveis Gharan AU - Alireza Rezaei BT - 29th Annual Conference on Learning Theory DA - 2016/06/06 ED - Vitaly Feldman ED - Alexander Rakhlin ED - Ohad Shamir ID - pmlr-v49-anari16 PB - PMLR DP - Proceedings of Machine Learning Research VL - 49 SP - 103 EP - 115 L1 - http://proceedings.mlr.press/v49/anari16.pdf UR - https://proceedings.mlr.press/v49/anari16.html AB - Strongly Rayleigh distributions are natural generalizations of product and determinantal probability distributions and satisfy the strongest form of negative dependence properties. We show that the "natural" Monte Carlo Markov Chain (MCMC) algorithm mixes rapidly in the support of a homogeneous strongly Rayleigh distribution. As a byproduct, our proof implies Markov chains can be used to efficiently generate approximate samples of a k-determinantal point process. This answers an open question raised by Deshpande and Rademacher which was studied recently by Kang, Li-Jegelka-Sra, and Rebeschini-Karbasi. ER -
APA
Anari, N., Oveis Gharan, S. & Rezaei, A.. (2016). Monte Carlo Markov Chain Algorithms for Sampling Strongly Rayleigh Distributions and Determinantal Point Processes. 29th Annual Conference on Learning Theory, in Proceedings of Machine Learning Research 49:103-115 Available from https://proceedings.mlr.press/v49/anari16.html.

Related Material