{"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/autosense-model-for-word-sense-induction","title":"AutoSense Model for Word Sense Induction","arxiv_id":"1811.09242","date":"2018-11-22","proceeding":null,"authors":["Reinald Kim Amplayo","Seung-won Hwang","Min Song"],"abstract":"Word sense induction (WSI), or the task of automatically discovering multiple\nsenses or meanings of a word, has three main challenges: domain adaptability,\nnovel sense detection, and sense granularity flexibility. While current latent\nvariable models are known to solve the first two challenges, they are not\nflexible to different word sense granularities, which differ very much among\nwords, from aardvark with one sense, to play with over 50 senses. Current\nmodels either require hyperparameter tuning or nonparametric induction of the\nnumber of senses, which we find both to be ineffective. Thus, we aim to\neliminate these requirements and solve the sense granularity problem by\nproposing AutoSense, a latent variable model based on two observations: (1)\nsenses are represented as a distribution over topics, and (2) senses generate\npairings between the target word and its neighboring word. These observations\nalleviate the problem by (a) throwing garbage senses and (b) additionally\ninducing fine-grained word senses. Results show great improvements over the\nstate-of-the-art models on popular WSI datasets. We also show that AutoSense is\nable to learn the appropriate sense granularity of a word. Finally, we apply\nAutoSense to the unsupervised author name disambiguation task where the sense\ngranularity problem is more evident and show that AutoSense is evidently better\nthan competing models. We share our data and code here:\nhttps://github.com/rktamplayo/AutoSense.","url_abs":"http://arxiv.org/abs/1811.09242v1","url_pdf":"http://arxiv.org/pdf/1811.09242v1.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":"autosense-model-for-word-sense-induction","repo_url":"https://github.com/rktamplayo/AutoSense","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"none","reach":null}],"tasks":[{"task_slug":"word-sense-induction","task_name":"Word Sense Induction"},{"task_slug":"model","task_name":"model"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/word-sense-induction-on-semeval-2010-wsi-1","task":"Word Sense Induction","dataset":"SemEval 2010 WSI","model":"AutoSense","rank_in_archive_order":2,"of":5,"metrics":{"AVG":"24.59","F-Score":"61.7","V-Measure":"9.8"},"uses_additional_data":false}],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}