{"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/challenges-in-disentangling-independent","title":"Challenges in Disentangling Independent Factors of Variation","arxiv_id":"1711.02245","date":"2017-11-07","proceeding":"ICLR 2018 1","authors":["Attila Szabó","Qiyang Hu","Tiziano Portenier","Matthias Zwicker","Paolo Favaro"],"abstract":"We study the problem of building models that disentangle independent factors\nof variation. Such models could be used to encode features that can efficiently\nbe used for classification and to transfer attributes between different images\nin image synthesis. As data we use a weakly labeled training set. Our weak\nlabels indicate what single factor has changed between two data samples,\nalthough the relative value of the change is unknown. This labeling is of\nparticular interest as it may be readily available without annotation costs. To\nmake use of weak labels we introduce an autoencoder model and train it through\nconstraints on image pairs and triplets. We formally prove that without\nadditional knowledge there is no guarantee that two images with the same factor\nof variation will be mapped to the same feature. We call this issue the\nreference ambiguity. Moreover, we show the role of the feature dimensionality\nand adversarial training. We demonstrate experimentally that the proposed model\ncan successfully transfer attributes on several datasets, but show also cases\nwhen the reference ambiguity occurs.","url_abs":"http://arxiv.org/abs/1711.02245v1","url_pdf":"http://arxiv.org/pdf/1711.02245v1.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":"challenges-in-disentangling-independent","repo_url":"https://github.com/ananyahjha93/challenges-in-disentangling","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}},{"paper_slug":"challenges-in-disentangling-independent","repo_url":"https://github.com/ananyahjha93/cycle-consistent-vae","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[{"task_slug":"image-generation","task_name":"Image Generation"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1711.02245","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}