{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/stochastic-bouncy-particle-sampler","title":"Stochastic Bouncy Particle Sampler","arxiv_id":"1609.00770","date":"2016-09-03","proceeding":"ICML 2017 8","authors":["Ari Pakman","Dar Gilboa","David Carlson","Liam Paninski"],"abstract":"We introduce a novel stochastic version of the non-reversible, rejection-free\nBouncy Particle Sampler (BPS), a Markov process whose sample trajectories are\npiecewise linear. The algorithm is based on simulating first arrival times in a\ndoubly stochastic Poisson process using the thinning method, and allows\nefficient sampling of Bayesian posteriors in big datasets. We prove that in the\nBPS no bias is introduced by noisy evaluations of the log-likelihood gradient.\nOn the other hand, we argue that efficiency considerations favor a small,\ncontrollable bias in the construction of the thinning proposals, in exchange\nfor faster mixing. We introduce a simple regression-based proposal intensity\nfor the thinning method that controls this trade-off. We illustrate the\nalgorithm in several examples in which it outperforms both unbiased, but slowly\nmixing stochastic versions of BPS, as well as biased stochastic gradient-based\nsamplers.","url_abs":"http://arxiv.org/abs/1609.00770v3","url_pdf":"http://arxiv.org/pdf/1609.00770v3.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"stochastic-bouncy-particle-sampler","repo_url":"https://github.com/dargilboa/SBPS-public","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"tf","reach":null}],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}