Improving Bengali and Hindi Large Language Models

Result code in IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F00216208%3A11320%2F25%3AMKDEC429" target="_blank" >RIV/00216208:11320/25:MKDEC429 - isvavai.cz</a>

Result on the web

<a href="https://www.scopus.com/inward/record.uri?eid=2-s2.0-85195934786&partnerID=40&md5=2c4071b1929b49b2e2f118841f834eca" target="_blank" >https://www.scopus.com/inward/record.uri?eid=2-s2.0-85195934786&partnerID=40&md5=2c4071b1929b49b2e2f118841f834eca</a>

DOI - Digital Object Identifier

—

Result language

angličtina

Original language name

Improving Bengali and Hindi Large Language Models

Original language description

Despite being widely spoken worldwide, Bengali and Hindi are low-resource languages. The state-of-the-art in modeling such languages uses BERT and the Wordpiece tokenizer. We observed that the Wordpiece tokenizer often breaks words into meaningless tokens, failing to separate roots from affixes. Moreover, Wordpiece does not take into account fine-grained character-level information. We hypothesize that modeling fine-grained character-level information or interactions between roots and affixes helps with modeling highly inflected and morphologically complex languages such as Bengali and Hindi. We used BERT with two different tokenizers - a Unigram tokenizer and a character-level tokenizer and observed better performance. Then, we pretrained four language models accordingly - Bengali Unigram BERT, Hindi Unigram BERT, Bengali Character BERT, and Hindi Character BERT, and evaluated them for masked token detection, both in correct and erroneous settings, across many NLU tasks. We provide experimental evidence that Unigram and character-level tokenizers lead to better pretrained models for Bengali and Hindi, outperforming the previous state-of-the-art and BERT with Wordpiece vocabulary. We conduct the first study investigating the efficacy of different tokenization methods in modeling Bengali and Hindi. © 2024 ELRA Language Resource Association: CC BY-NC 4.0.

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

Jt. Int. Conf. Comput. Linguist., Lang. Resour. Eval., LREC-COLING - Main Conf. Proc.

ISBN

978-249381410-4

ISSN

—

e-ISSN

—

Number of pages

13

Pages from-to

8719-8731

Publisher name

European Language Resources Association (ELRA)

Place of publication

—

Event location

Torino, Italia

Event date

Jan 1, 2025

Type of event by nationality

WRD - Celosvětová akce

UT code for WoS article

—