{"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/neural-symbolic-machines-learning-semantic-1","title":"Neural Symbolic Machines: Learning Semantic Parsers on Freebase with Weak Supervision","arxiv_id":"1611.00020","date":"2016-10-31","proceeding":"ACL 2017 7","authors":["Chen Liang","Jonathan Berant","Quoc Le","Kenneth D. Forbus","Ni Lao"],"abstract":"Harnessing the statistical power of neural networks to perform language\nunderstanding and symbolic reasoning is difficult, when it requires executing\nefficient discrete operations against a large knowledge-base. In this work, we\nintroduce a Neural Symbolic Machine, which contains (a) a neural \"programmer\",\ni.e., a sequence-to-sequence model that maps language utterances to programs\nand utilizes a key-variable memory to handle compositionality (b) a symbolic\n\"computer\", i.e., a Lisp interpreter that performs program execution, and helps\nfind good programs by pruning the search space. We apply REINFORCE to directly\noptimize the task reward of this structured prediction problem. To train with\nweak supervision and improve the stability of REINFORCE, we augment it with an\niterative maximum-likelihood training process. NSM outperforms the\nstate-of-the-art on the WebQuestionsSP dataset when trained from\nquestion-answer pairs only, without requiring any feature engineering or\ndomain-specific knowledge.","url_abs":"http://arxiv.org/abs/1611.00020v4","url_pdf":"http://arxiv.org/pdf/1611.00020v4.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":"neural-symbolic-machines-learning-semantic-1","repo_url":"https://github.com/crazydonkey200/neural-symbolic-machines","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok","spdx":"Apache-2.0"}},{"paper_slug":"neural-symbolic-machines-learning-semantic-1","repo_url":"https://github.com/theSparta/neural-symbolic-machines","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok","spdx":"Apache-2.0"}}],"tasks":[{"task_slug":"feature-engineering","task_name":"Feature Engineering"},{"task_slug":"structured-prediction","task_name":"Structured Prediction"}],"methods":[{"method_slug":"pruning","method_name":"Pruning"},{"method_slug":"reinforce","method_name":"REINFORCE"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1611.00020","atlas_url":"https://app.syntology.ai/?focus=1611.00020","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}