{"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/conditional-inference-in-pre-trained","title":"Conditional Inference in Pre-trained Variational Autoencoders via Cross-coding","arxiv_id":"1805.07785","date":"2018-05-20","proceeding":"ICLR 2019 5","authors":["Ga Wu","Justin Domke","Scott Sanner"],"abstract":"Variational Autoencoders (VAEs) are a popular generative model, but one in\nwhich conditional inference can be challenging. If the decomposition into query\nand evidence variables is fixed, conditional VAEs provide an attractive\nsolution. To support arbitrary queries, one is generally reduced to Markov\nChain Monte Carlo sampling methods that can suffer from long mixing times. In\nthis paper, we propose an idea we term cross-coding to approximate the\ndistribution over the latent variables after conditioning on an evidence\nassignment to some subset of the variables. This allows generating query\nsamples without retraining the full VAE. We experimentally evaluate three\nvariations of cross-coding showing that (i) they can be quickly optimized for\ndifferent decompositions of evidence and query and (ii) they quantitatively and\nqualitatively outperform Hamiltonian Monte Carlo.","url_abs":"http://arxiv.org/abs/1805.07785v2","url_pdf":"http://arxiv.org/pdf/1805.07785v2.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":"conditional-inference-in-pre-trained","repo_url":"https://github.com/wuga214/XCoder_VAE_Conditional_Inference","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"tf","reach":null}],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}