{"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/probabilistic-fasttext-for-multi-sense-word","title":"Probabilistic FastText for Multi-Sense Word Embeddings","arxiv_id":"1806.02901","date":"2018-06-07","proceeding":"ACL 2018 7","authors":["Ben Athiwaratkun","Andrew Gordon Wilson","Anima Anandkumar"],"abstract":"We introduce Probabilistic FastText, a new model for word embeddings that can\ncapture multiple word senses, sub-word structure, and uncertainty information.\nIn particular, we represent each word with a Gaussian mixture density, where\nthe mean of a mixture component is given by the sum of n-grams. This\nrepresentation allows the model to share statistical strength across sub-word\nstructures (e.g. Latin roots), producing accurate representations of rare,\nmisspelt, or even unseen words. Moreover, each component of the mixture can\ncapture a different word sense. Probabilistic FastText outperforms both\nFastText, which has no probabilistic model, and dictionary-level probabilistic\nembeddings, which do not incorporate subword structures, on several\nword-similarity benchmarks, including English RareWord and foreign language\ndatasets. We also achieve state-of-art performance on benchmarks that measure\nability to discern different meanings. Thus, the proposed model is the first to\nachieve multi-sense representations while having enriched semantics on rare\nwords.","url_abs":"http://arxiv.org/abs/1806.02901v1","url_pdf":"http://arxiv.org/pdf/1806.02901v1.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":"probabilistic-fasttext-for-multi-sense-word","repo_url":"https://github.com/benathi/multisense-prob-fasttext","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok","spdx":"NOASSERTION"}}],"tasks":[{"task_slug":"word-embeddings","task_name":"Word Embeddings"},{"task_slug":"word-similarity","task_name":"Word Similarity"}],"methods":[{"method_slug":"fasttext","method_name":"fastText"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1806.02901","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}