{"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/multi-level-variational-autoencoder-learning","title":"Multi-Level Variational Autoencoder: Learning Disentangled Representations from Grouped Observations","arxiv_id":"1705.08841","date":"2017-05-24","proceeding":null,"authors":["Diane Bouchacourt","Ryota Tomioka","Sebastian Nowozin"],"abstract":"We would like to learn a representation of the data which decomposes an\nobservation into factors of variation which we can independently control.\nSpecifically, we want to use minimal supervision to learn a latent\nrepresentation that reflects the semantics behind a specific grouping of the\ndata, where within a group the samples share a common factor of variation. For\nexample, consider a collection of face images grouped by identity. We wish to\nanchor the semantics of the grouping into a relevant and disentangled\nrepresentation that we can easily exploit. However, existing deep probabilistic\nmodels often assume that the observations are independent and identically\ndistributed. We present the Multi-Level Variational Autoencoder (ML-VAE), a new\ndeep probabilistic model for learning a disentangled representation of a set of\ngrouped observations. The ML-VAE separates the latent representation into\nsemantically meaningful parts by working both at the group level and the\nobservation level, while retaining efficient test-time inference. Quantitative\nand qualitative evaluations show that the ML-VAE model (i) learns a\nsemantically meaningful disentanglement of grouped data, (ii) enables\nmanipulation of the latent representation, and (iii) generalises to unseen\ngroups.","url_abs":"http://arxiv.org/abs/1705.08841v1","url_pdf":"http://arxiv.org/pdf/1705.08841v1.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":"multi-level-variational-autoencoder-learning","repo_url":"https://github.com/ananyahjha93/multi-level-vae","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"unanswered"}}],"tasks":[{"task_slug":"disentanglement","task_name":"Disentanglement"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1705.08841","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}