{"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/latent-predictor-networks-for-code-generation","title":"Latent Predictor Networks for Code Generation","arxiv_id":"1603.06744","date":"2016-03-22","proceeding":"ACL 2016 8","authors":["Wang Ling","Edward Grefenstette","Karl Moritz Hermann","Tomáš Kočiský","Andrew Senior","Fumin Wang","Phil Blunsom"],"abstract":"Many language generation tasks require the production of text conditioned on\nboth structured and unstructured inputs. We present a novel neural network\narchitecture which generates an output sequence conditioned on an arbitrary\nnumber of input functions. Crucially, our approach allows both the choice of\nconditioning context and the granularity of generation, for example characters\nor tokens, to be marginalised, thus permitting scalable and effective training.\nUsing this framework, we address the problem of generating programming code\nfrom a mixed natural language and structured specification. We create two new\ndata sets for this paradigm derived from the collectible trading card games\nMagic the Gathering and Hearthstone. On these, and a third preexisting corpus,\nwe demonstrate that marginalising multiple predictors allows our model to\noutperform strong benchmarks.","url_abs":"http://arxiv.org/abs/1603.06744v2","url_pdf":"http://arxiv.org/pdf/1603.06744v2.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":"latent-predictor-networks-for-code-generation","repo_url":"https://github.com/deepmind/card2code","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok","spdx":"NOASSERTION"}},{"paper_slug":"latent-predictor-networks-for-code-generation","repo_url":"https://github.com/davidgolub/QuestionGeneration","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"NOASSERTION"}}],"tasks":[{"task_slug":"card-games","task_name":"Card Games"},{"task_slug":"code-generation","task_name":"Code Generation"},{"task_slug":"text-generation","task_name":"Text Generation"}],"methods":[],"datasets_introduced":[{"slug":"hearthstone","name":"Hearthstone","full_name":""}],"methods_introduced":[],"results":[{"leaderboard":"/sota/code-generation-on-django","task":"Code Generation","dataset":"Django","model":"lpn (Ling et al., 2016)","rank_in_archive_order":10,"of":11,"metrics":{"Accuracy":"62.3","BLEU Score":"77.6"},"uses_additional_data":false},{"leaderboard":"/sota/code-generation-on-django","task":"Code Generation","dataset":"Django","model":"Phrasal Statistical MT (Ling et al., 2016)","rank_in_archive_order":11,"of":11,"metrics":{"Accuracy":"31.5","BLEU Score":"47.6"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1603.06744","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}