Carregant...
Fitxers
Tipus de document
ArticleVersió
Versió publicadaData de publicació
Llicència de publicació
Si us plau utilitzeu sempre aquest identificador per citar o enllaçar aquest document: https://hdl.handle.net/2445/192367
Information theory-based compositional distributional semantics
Títol de la revista
Director/Tutor
ISSN de la revista
Títol del volum
Recurs relacionat
Resum
In the context of text representation, Compositional Distributional Semantics models aim to fuse the Distributional Hypothesis and the Principle of Compositionality. Text embedding is based on co-ocurrence distributions and the representations are in turn combined by compositional functions taking into account the text structure. However, the theoretical basis of compositional functions is still an open issue. In this article we define and study the notion of Information Theory-based Compositional Distributional Semantics (ICDS): (i) We first establish formal properties for embedding, composition, and similarity functions based on Shannon's Information Theory; (ii) we analyze the existing approaches under this prism, checking whether or not they comply with the established desirable properties; (iii) we propose two parameterizable composition and similarity functions that generalize traditional approaches while fulfilling the formal properties; and finally (iv) we perform an empirical study on several textual similarity datasets that include sentences with a high and low lexical overlap, and on the similarity between words and their description. Our theoretical analysis and empirical results show that fulfilling formal properties affects positively the accuracy of text representation models in terms of correspondence (isometry) between the embedding and meaning spaces.
Matèries
Matèries (anglès)
Citació
Citació
AMIGÓ, Enrique, ARIZA CASABONA, Alejandro, FRESNOS, Víctor, MARTÍ ANTONIN, M. antònia. Information theory-based compositional distributional semantics. _Computational Linguistics_. 2022. Vol. 48, núm. 4, pàgs. 907-948. [consulta: 20 de gener de 2026]. ISSN: 0891-2017. [Disponible a: https://hdl.handle.net/2445/192367]