{"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/piecewise-latent-variables-for-neural","title":"Piecewise Latent Variables for Neural Variational Text Processing","arxiv_id":"1612.00377","date":"2016-12-01","proceeding":"EMNLP (ACL) 2017 9","authors":["Iulian V. Serban","Alexander G. Ororbia II","Joelle Pineau","Aaron Courville"],"abstract":"Advances in neural variational inference have facilitated the learning of\npowerful directed graphical models with continuous latent variables, such as\nvariational autoencoders. The hope is that such models will learn to represent\nrich, multi-modal latent factors in real-world data, such as natural language\ntext. However, current models often assume simplistic priors on the latent\nvariables - such as the uni-modal Gaussian distribution - which are incapable\nof representing complex latent factors efficiently. To overcome this\nrestriction, we propose the simple, but highly flexible, piecewise constant\ndistribution. This distribution has the capacity to represent an exponential\nnumber of modes of a latent target distribution, while remaining mathematically\ntractable. Our results demonstrate that incorporating this new latent\ndistribution into different models yields substantial improvements in natural\nlanguage processing tasks such as document modeling and natural language\ngeneration for dialogue.","url_abs":"http://arxiv.org/abs/1612.00377v4","url_pdf":"http://arxiv.org/pdf/1612.00377v4.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":"piecewise-latent-variables-for-neural","repo_url":"https://github.com/julianser/hred-latent-piecewise","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"none","reach":null},{"paper_slug":"piecewise-latent-variables-for-neural","repo_url":"https://github.com/ago109/piecewise-nvdm-emnlp-2017","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"none","reach":null}],"tasks":[{"task_slug":"text-generation","task_name":"Text Generation"},{"task_slug":"variational-inference","task_name":"Variational Inference"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}