{"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/beyond-expectation-deep-joint-mean-and","title":"Beyond expectation: Deep joint mean and quantile regression for spatio-temporal problems","arxiv_id":"1808.08798","date":"2018-08-27","proceeding":null,"authors":["Filipe Rodrigues","Francisco C. Pereira"],"abstract":"Spatio-temporal problems are ubiquitous and of vital importance in many\nresearch fields. Despite the potential already demonstrated by deep learning\nmethods in modeling spatio-temporal data, typical approaches tend to focus\nsolely on conditional expectations of the output variables being modeled. In\nthis paper, we propose a multi-output multi-quantile deep learning approach for\njointly modeling several conditional quantiles together with the conditional\nexpectation as a way to provide a more complete \"picture\" of the predictive\ndensity in spatio-temporal problems. Using two large-scale datasets from the\ntransportation domain, we empirically demonstrate that, by approaching the\nquantile regression problem from a multi-task learning perspective, it is\npossible to solve the embarrassing quantile crossings problem, while\nsimultaneously significantly outperforming state-of-the-art quantile regression\nmethods. Moreover, we show that jointly modeling the mean and several\nconditional quantiles not only provides a rich description about the predictive\ndensity that can capture heteroscedastic properties at a neglectable\ncomputational overhead, but also leads to improved predictions of the\nconditional expectation due to the extra information and a regularization\neffect induced by the added quantiles.","url_abs":"http://arxiv.org/abs/1808.08798v1","url_pdf":"http://arxiv.org/pdf/1808.08798v1.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":"beyond-expectation-deep-joint-mean-and","repo_url":"https://github.com/fmpr/DeepJMQR","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok","spdx":"GPL-3.0"}}],"tasks":[{"task_slug":"multi-task-learning","task_name":"Multi-Task Learning"},{"task_slug":"quantile-regression","task_name":"quantile regression"},{"task_slug":"regression-1","task_name":"regression"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1808.08798","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}