{"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/numeracy-for-language-models-evaluating-and","title":"Numeracy for Language Models: Evaluating and Improving their Ability to Predict Numbers","arxiv_id":"1805.08154","date":"2018-05-21","proceeding":"ACL 2018 7","authors":["Georgios P. Spithourakis","Sebastian Riedel"],"abstract":"Numeracy is the ability to understand and work with numbers. It is a\nnecessary skill for composing and understanding documents in clinical,\nscientific, and other technical domains. In this paper, we explore different\nstrategies for modelling numerals with language models, such as memorisation\nand digit-by-digit composition, and propose a novel neural architecture that\nuses a continuous probability density function to model numerals from an open\nvocabulary. Our evaluation on clinical and scientific datasets shows that using\nhierarchical models to distinguish numerals from words improves a perplexity\nmetric on the subset of numerals by 2 and 4 orders of magnitude, respectively,\nover non-hierarchical models. A combination of strategies can further improve\nperplexity. Our continuous probability density function model reduces mean\nabsolute percentage errors by 18% and 54% in comparison to the second best\nstrategy for each dataset, respectively.","url_abs":"http://arxiv.org/abs/1805.08154v1","url_pdf":"http://arxiv.org/pdf/1805.08154v1.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":"numeracy-for-language-models-evaluating-and","repo_url":"https://github.com/uclmr/numerate-language-models","is_official":0,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"none","reach":{"status":"ok","spdx":"GPL-3.0"}}],"tasks":[{"task_slug":"language-modelling","task_name":"Language Modelling"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1805.08154","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}