{"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/large-scale-stochastic-sampling-from-the","title":"Large-Scale Stochastic Sampling from the Probability Simplex","arxiv_id":"1806.07137","date":"2018-06-19","proceeding":"NeurIPS 2018 12","authors":["Jack Baker","Paul Fearnhead","Emily B. Fox","Christopher Nemeth"],"abstract":"Stochastic gradient Markov chain Monte Carlo (SGMCMC) has become a popular\nmethod for scalable Bayesian inference. These methods are based on sampling a\ndiscrete-time approximation to a continuous time process, such as the Langevin\ndiffusion. When applied to distributions defined on a constrained space the\ntime-discretization error can dominate when we are near the boundary of the\nspace. We demonstrate that because of this, current SGMCMC methods for the\nsimplex struggle with sparse simplex spaces; when many of the components are\nclose to zero. Unfortunately, many popular large-scale Bayesian models, such as\nnetwork or topic models, require inference on sparse simplex spaces. To avoid\nthe biases caused by this discretization error, we propose the stochastic\nCox-Ingersoll-Ross process (SCIR), which removes all discretization error and\nwe prove that samples from the SCIR process are asymptotically unbiased. We\ndiscuss how this idea can be extended to target other constrained spaces. Use\nof the SCIR process within a SGMCMC algorithm is shown to give substantially\nbetter performance for a topic model and a Dirichlet process mixture model than\nexisting SGMCMC approaches.","url_abs":"http://arxiv.org/abs/1806.07137v2","url_pdf":"http://arxiv.org/pdf/1806.07137v2.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":"large-scale-stochastic-sampling-from-the","repo_url":"https://github.com/jbaker92/scir","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"none","reach":null}],"tasks":[{"task_slug":"bayesian-inference","task_name":"Bayesian Inference"},{"task_slug":"topic-models","task_name":"Topic Models"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}