{"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/distinct-patterns-of-syntactic-agreement","title":"Distinct patterns of syntactic agreement errors in recurrent networks and humans","arxiv_id":"1807.06882","date":"2018-07-18","proceeding":null,"authors":["Tal Linzen","Brian Leonard"],"abstract":"Determining the correct form of a verb in context requires an understanding\nof the syntactic structure of the sentence. Recurrent neural networks have been\nshown to perform this task with an error rate comparable to humans, despite the\nfact that they are not designed with explicit syntactic representations. To\nexamine the extent to which the syntactic representations of these networks are\nsimilar to those used by humans when processing sentences, we compare the\ndetailed pattern of errors that RNNs and humans make on this task. Despite\nsignificant similarities (attraction errors, asymmetry between singular and\nplural subjects), the error patterns differed in important ways. In particular,\nin complex sentences with relative clauses error rates increased in RNNs but\ndecreased in humans. Furthermore, RNNs showed a cumulative effect of attractors\nbut humans did not. We conclude that at least in some respects the syntactic\nrepresentations acquired by RNNs are fundamentally different from those used by\nhumans.","url_abs":"http://arxiv.org/abs/1807.06882v1","url_pdf":"http://arxiv.org/pdf/1807.06882v1.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":"distinct-patterns-of-syntactic-agreement","repo_url":"https://github.com/jhupsycholing/RNNvsHumanSyntax","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"none","reach":null}],"tasks":[{"task_slug":"sentence","task_name":"Sentence"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1807.06882","atlas_url":"https://app.syntology.ai/?focus=1807.06882","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}