Navigating Cross-Lingual Natural Language Processing: Challenges, Strategies, and Applications

Result code in IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F00216208%3A11320%2F25%3AI89URPFH" target="_blank" >RIV/00216208:11320/25:I89URPFH - isvavai.cz</a>

Result on the web

<a href="https://www.scopus.com/inward/record.uri?eid=2-s2.0-85201101707&doi=10.1007%2f978-981-97-2716-2_19&partnerID=40&md5=39ce5825df4f63ba7fdd524197d8fdaf" target="_blank" >https://www.scopus.com/inward/record.uri?eid=2-s2.0-85201101707&doi=10.1007%2f978-981-97-2716-2_19&partnerID=40&md5=39ce5825df4f63ba7fdd524197d8fdaf</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1007/978-981-97-2716-2_19" target="_blank" >10.1007/978-981-97-2716-2_19</a>

Result language

angličtina

Original language name

Navigating Cross-Lingual Natural Language Processing: Challenges, Strategies, and Applications

Original language description

Modern techniques for most Natural Language Processing (NLP) activities have attained near-human functionality. This latest advancement has benefited countless individuals and companies throughout the globe. Unfortunately, most massive tagged databases are only accessible in a couple of languages; for many languages, either few or no tags are accessible to enable automatic NLP uses. As a result, among the priorities of cross-lingual NLP study is to build computing algorithms that leverage abundant resource corpora of languages and employ them in limited-resource language uses through transportable illustration learning. The paper covers the basic difficulties and suggests multiple approaches for cross-lingual illustration learning that employ common syntax dependence to link typological variations across languages and efficiently use unmarked supplies to learn solid and generalizable depictions. The methodologies suggested in this research efficiently translate across a broad spectrum of languages and NLP uses such as dependent parsing, titled entity identification, text categorization, query replying, and others. Test outcomes reveal that enhancing mBERT with syntax increases cross-lingual transfer by 1.4 and 1.6 scores on average for every targeted language in PAWS-X and MLQR, respectively. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024.

Czech name

—

Czech description

—

Type

D - Article in proceedings

CEP classification

—

OECD FORD branch

10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Project

—

Continuities

—

Publication year

2024

Confidentiality

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Article name in the collection

Smart Innov. Syst. Technol.

ISBN

978-981972715-5

ISSN

2190-3018

e-ISSN

—

Number of pages

12

Pages from-to

203-214

Publisher name

Springer Science and Business Media Deutschland GmbH

Place of publication

—

Event location

Noida

Event date

Jan 1, 2025

Type of event by nationality

WRD - Celosvětová akce

UT code for WoS article

—