{"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/fixing-a-broken-elbo","title":"Fixing a Broken ELBO","arxiv_id":"1711.00464","date":"2017-11-01","proceeding":"ICML 2018 7","authors":["Alexander A. Alemi","Ben Poole","Ian Fischer","Joshua V. Dillon","Rif A. Saurous","Kevin Murphy"],"abstract":"Recent work in unsupervised representation learning has focused on learning\ndeep directed latent-variable models. Fitting these models by maximizing the\nmarginal likelihood or evidence is typically intractable, thus a common\napproximation is to maximize the evidence lower bound (ELBO) instead. However,\nmaximum likelihood training (whether exact or approximate) does not necessarily\nresult in a good latent representation, as we demonstrate both theoretically\nand empirically. In particular, we derive variational lower and upper bounds on\nthe mutual information between the input and the latent variable, and use these\nbounds to derive a rate-distortion curve that characterizes the tradeoff\nbetween compression and reconstruction accuracy. Using this framework, we\ndemonstrate that there is a family of models with identical ELBO, but different\nquantitative and qualitative characteristics. Our framework also suggests a\nsimple new method to ensure that latent variable models with powerful\nstochastic decoders do not ignore their latent code.","url_abs":"http://arxiv.org/abs/1711.00464v3","url_pdf":"http://arxiv.org/pdf/1711.00464v3.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":"fixing-a-broken-elbo","repo_url":"https://github.com/suvalaki/Deeper","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok","spdx":"GPL-3.0"}}],"tasks":[{"task_slug":"representation-learning","task_name":"Representation Learning"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1711.00464","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}