Incorporating Word Sense Disambiguation in Neural Language Models

Incorporating Word Sense Disambiguation in Neural Language Models. Wahle, J. P., Ruas, T., Meuschke, N., & Gipp, B. arXiv:2106.07967 [cs], June, 2021. arXiv: 2106.07967

Paper abstract bibtex

We present two supervised (pre-)training methods to incorporate gloss definitions from lexical resources into neural language models (LMs). The training improves our models' performance for Word Sense Disambiguation (WSD) but also benefits general language understanding tasks while adding almost no parameters. We evaluate our techniques with seven different neural LMs and find that XLNet is more suitable for WSD than BERT. Our best-performing methods exceeds state-of-the-art WSD techniques on the SemCor 3.0 dataset by 0.5% F1 and increase BERT's performance on the GLUE benchmark by 1.1% on average.

@article{WahleRMG21a,
	title = {Incorporating {Word} {Sense} {Disambiguation} in {Neural} {Language} {Models}},
	url = {https://arxiv.org/pdf/2106.07967.pdf},
	abstract = {We present two supervised (pre-)training methods to incorporate gloss definitions from lexical resources into neural language models (LMs). The training improves our models' performance for Word Sense Disambiguation (WSD) but also benefits general language understanding tasks while adding almost no parameters. We evaluate our techniques with seven different neural LMs and find that XLNet is more suitable for WSD than BERT. Our best-performing methods exceeds state-of-the-art WSD techniques on the SemCor 3.0 dataset by 0.5\% F1 and increase BERT's performance on the GLUE benchmark by 1.1\% on average.},
	urldate = {2021-06-19},
	journal = {arXiv:2106.07967 [cs]},
	author = {Wahle, Jan Philip and Ruas, Terry and Meuschke, Norman and Gipp, Bela},
	month = jun,
	year = {2021},
	note = {arXiv: 2106.07967},
	keywords = {!tr\_author, Computer Science - Artificial Intelligence, Computer Science - Computation and Language, \_pre\_print, lang\_model, nlp, wsd, ⛔ No DOI found},
}

Downloads: 0

{"_id":"w5pMHhJntFuW5avap","bibbaseid":"wahle-ruas-meuschke-gipp-incorporatingwordsensedisambiguationinneurallanguagemodels-2021","author_short":["Wahle, J. P.","Ruas, T.","Meuschke, N.","Gipp, B."],"bibdata":{"bibtype":"article","type":"article","title":"Incorporating Word Sense Disambiguation in Neural Language Models","url":"https://arxiv.org/pdf/2106.07967.pdf","abstract":"We present two supervised (pre-)training methods to incorporate gloss definitions from lexical resources into neural language models (LMs). The training improves our models' performance for Word Sense Disambiguation (WSD) but also benefits general language understanding tasks while adding almost no parameters. We evaluate our techniques with seven different neural LMs and find that XLNet is more suitable for WSD than BERT. Our best-performing methods exceeds state-of-the-art WSD techniques on the SemCor 3.0 dataset by 0.5% F1 and increase BERT's performance on the GLUE benchmark by 1.1% on average.","urldate":"2021-06-19","journal":"arXiv:2106.07967 [cs]","author":[{"propositions":[],"lastnames":["Wahle"],"firstnames":["Jan","Philip"],"suffixes":[]},{"propositions":[],"lastnames":["Ruas"],"firstnames":["Terry"],"suffixes":[]},{"propositions":[],"lastnames":["Meuschke"],"firstnames":["Norman"],"suffixes":[]},{"propositions":[],"lastnames":["Gipp"],"firstnames":["Bela"],"suffixes":[]}],"month":"June","year":"2021","note":"arXiv: 2106.07967","keywords":"!tr_author, Computer Science - Artificial Intelligence, Computer Science - Computation and Language, _pre_print, lang_model, nlp, wsd, ⛔ No DOI found","bibtex":"@article{WahleRMG21a,\n\ttitle = {Incorporating {Word} {Sense} {Disambiguation} in {Neural} {Language} {Models}},\n\turl = {https://arxiv.org/pdf/2106.07967.pdf},\n\tabstract = {We present two supervised (pre-)training methods to incorporate gloss definitions from lexical resources into neural language models (LMs). The training improves our models' performance for Word Sense Disambiguation (WSD) but also benefits general language understanding tasks while adding almost no parameters. We evaluate our techniques with seven different neural LMs and find that XLNet is more suitable for WSD than BERT. Our best-performing methods exceeds state-of-the-art WSD techniques on the SemCor 3.0 dataset by 0.5\\% F1 and increase BERT's performance on the GLUE benchmark by 1.1\\% on average.},\n\turldate = {2021-06-19},\n\tjournal = {arXiv:2106.07967 [cs]},\n\tauthor = {Wahle, Jan Philip and Ruas, Terry and Meuschke, Norman and Gipp, Bela},\n\tmonth = jun,\n\tyear = {2021},\n\tnote = {arXiv: 2106.07967},\n\tkeywords = {!tr\\_author, Computer Science - Artificial Intelligence, Computer Science - Computation and Language, \\_pre\\_print, lang\\_model, nlp, wsd, ⛔ No DOI found},\n}\n","author_short":["Wahle, J. P.","Ruas, T.","Meuschke, N.","Gipp, B."],"key":"WahleRMG21a","id":"WahleRMG21a","bibbaseid":"wahle-ruas-meuschke-gipp-incorporatingwordsensedisambiguationinneurallanguagemodels-2021","role":"author","urls":{"Paper":"https://arxiv.org/pdf/2106.07967.pdf"},"keyword":["!tr_author","Computer Science - Artificial Intelligence","Computer Science - Computation and Language","_pre_print","lang_model","nlp","wsd","⛔ No DOI found"],"metadata":{"authorlinks":{}}},"bibtype":"article","biburl":"https://api.zotero.org/groups/2503580/items?key=9bMVo5bWhsSJ7a6YWgBxjXpk&format=bibtex&limit=100","dataSources":["XJBi8b8xDjDoWPzcZ","hG7rv86o2PDG2z44d"],"keywords":["!tr_author","computer science - artificial intelligence","computer science - computation and language","_pre_print","lang_model","nlp","wsd","⛔ no doi found"],"search_terms":["incorporating","word","sense","disambiguation","neural","language","models","wahle","ruas","meuschke","gipp"],"title":"Incorporating Word Sense Disambiguation in Neural Language Models","year":2021,"downloads":1}