{"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/combining-symbolic-expressions-and-black-box","title":"Combining Symbolic Expressions and Black-box Function Evaluations in Neural Programs","arxiv_id":"1801.04342","date":"2018-01-12","proceeding":"ICLR 2018 1","authors":["Forough Arabshahi","Sameer Singh","Animashree Anandkumar"],"abstract":"Neural programming involves training neural networks to learn programs,\nmathematics, or logic from data. Previous works have failed to achieve good\ngeneralization performance, especially on problems and programs with high\ncomplexity or on large domains. This is because they mostly rely either on\nblack-box function evaluations that do not capture the structure of the\nprogram, or on detailed execution traces that are expensive to obtain, and\nhence the training data has poor coverage of the domain under consideration. We\npresent a novel framework that utilizes black-box function evaluations, in\nconjunction with symbolic expressions that define relationships between the\ngiven functions. We employ tree LSTMs to incorporate the structure of the\nsymbolic expression trees. We use tree encoding for numbers present in function\nevaluation data, based on their decimal representation. We present an\nevaluation benchmark for this task to demonstrate our proposed model combines\nsymbolic reasoning and function evaluation in a fruitful manner, obtaining high\naccuracies in our experiments. Our framework generalizes significantly better\nto expressions of higher depth and is able to fill partial equations with valid\ncompletions.","url_abs":"http://arxiv.org/abs/1801.04342v3","url_pdf":"http://arxiv.org/pdf/1801.04342v3.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":"combining-symbolic-expressions-and-black-box","repo_url":"https://github.com/ForoughA/neuralMath","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"mxnet","reach":{"status":"ok"}}],"tasks":[{"task_slug":null,"task_name":"valid"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":"https://app.syntology.ai/?focus=1801.04342","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}