A Morphology-Based Investigation of Positional Encodings

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Ghosh, Poulami, Vashishth, Shikhar, Dabre, Raj, Bhattacharyya, Pushpak
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866914816265814016
author Ghosh, Poulami
Vashishth, Shikhar
Dabre, Raj
Bhattacharyya, Pushpak
author_facet Ghosh, Poulami
Vashishth, Shikhar
Dabre, Raj
Bhattacharyya, Pushpak
contents Contemporary deep learning models effectively handle languages with diverse morphology despite not being directly integrated into them. Morphology and word order are closely linked, with the latter incorporated into transformer-based models through positional encodings. This prompts a fundamental inquiry: Is there a correlation between the morphological complexity of a language and the utilization of positional encoding in pre-trained language models? In pursuit of an answer, we present the first study addressing this question, encompassing 22 languages and 5 downstream tasks. Our findings reveal that the importance of positional encoding diminishes with increasing morphological complexity in languages. Our study motivates the need for a deeper understanding of positional encoding, augmenting them to better reflect the different languages under consideration.
format Preprint
id arxiv_https___arxiv_org_abs_2404_04530
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle A Morphology-Based Investigation of Positional Encodings
Ghosh, Poulami
Vashishth, Shikhar
Dabre, Raj
Bhattacharyya, Pushpak
Computation and Language
Contemporary deep learning models effectively handle languages with diverse morphology despite not being directly integrated into them. Morphology and word order are closely linked, with the latter incorporated into transformer-based models through positional encodings. This prompts a fundamental inquiry: Is there a correlation between the morphological complexity of a language and the utilization of positional encoding in pre-trained language models? In pursuit of an answer, we present the first study addressing this question, encompassing 22 languages and 5 downstream tasks. Our findings reveal that the importance of positional encoding diminishes with increasing morphological complexity in languages. Our study motivates the need for a deeper understanding of positional encoding, augmenting them to better reflect the different languages under consideration.
title A Morphology-Based Investigation of Positional Encodings
topic Computation and Language
url https://arxiv.org/abs/2404.04530