{"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/learning-a-generative-model-for-validity-in","title":"Learning a Generative Model for Validity in Complex Discrete Structures","arxiv_id":"1712.01664","date":"2017-12-05","proceeding":"ICLR 2018 1","authors":["David Janz","Jos van der Westhuizen","Brooks Paige","Matt J. Kusner","José Miguel Hernández-Lobato"],"abstract":"Deep generative models have been successfully used to learn representations\nfor high-dimensional discrete spaces by representing discrete objects as\nsequences and employing powerful sequence-based deep models. Unfortunately,\nthese sequence-based models often produce invalid sequences: sequences which do\nnot represent any underlying discrete structure; invalid sequences hinder the\nutility of such models. As a step towards solving this problem, we propose to\nlearn a deep recurrent validator model, which can estimate whether a partial\nsequence can function as the beginning of a full, valid sequence. This\nvalidator provides insight as to how individual sequence elements influence the\nvalidity of the overall sequence, and can be used to constrain sequence based\nmodels to generate valid sequences -- and thus faithfully model discrete\nobjects. Our approach is inspired by reinforcement learning, where an oracle\nwhich can evaluate validity of complete sequences provides a sparse reward\nsignal. We demonstrate its effectiveness as a generative model of Python 3\nsource code for mathematical expressions, and in improving the ability of a\nvariational autoencoder trained on SMILES strings to decode valid molecular\nstructures.","url_abs":"http://arxiv.org/abs/1712.01664v4","url_pdf":"http://arxiv.org/pdf/1712.01664v4.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":"learning-a-generative-model-for-validity-in","repo_url":"https://github.com/DavidJanz/molecule_grammar_rnn","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"reinforcement-learning","task_name":"Reinforcement Learning"},{"task_slug":null,"task_name":"valid"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}