We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.CL

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Computation and Language

Title: Language models in word sense disambiguation for Polish

Abstract: In the paper, we test two different approaches to the {unsupervised} word sense disambiguation task for Polish. In both methods, we use neural language models to predict words similar to those being disambiguated and, on the basis of these words, we predict the partition of word senses in different ways. In the first method, we cluster selected similar words, while in the second, we cluster vectors representing their subsets. The evaluation was carried out on texts annotated with plWordNet senses and provided a relatively good result (F1=0.68 for all ambiguous words). The results are significantly better than those obtained for the neural model-based unsupervised method proposed in \cite{waw:myk:17:Sense} and are at the level of the supervised method presented there. The proposed method may be a way of solving word sense disambiguation problem for languages that lack sense annotated data.
Subjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
Cite as: arXiv:2111.13982 [cs.CL]
  (or arXiv:2111.13982v1 [cs.CL] for this version)

Submission history

From: Agnieszka Mykowiecka [view email]
[v1] Sat, 27 Nov 2021 20:47:53 GMT (38kb)

Link back to: arXiv, form interface, contact.