CNN for Modeling Sanskrit Originated Bengali and Hindi Language

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F00216208%3A11320%2F22%3AN4K5R3E8" target="_blank" >RIV/00216208:11320/22:N4K5R3E8 - isvavai.cz</a>

Výsledek na webu

<a href="https://aclanthology.org/2022.aacl-main.4" target="_blank" >https://aclanthology.org/2022.aacl-main.4</a>

DOI - Digital Object Identifier

—

Jazyk výsledku

angličtina

Název v původním jazyce

CNN for Modeling Sanskrit Originated Bengali and Hindi Language

Popis výsledku v původním jazyce

Though recent works have focused on modeling high resource languages, the area is still unexplored for low resource languages like Bengali and Hindi. We propose an end to end trainable memory efficient CNN architecture named CoCNN to handle specific characteristics such as high inflection, morphological richness, flexible word order and phonetical spelling errors of Bengali and Hindi. In particular, we introduce two learnable convolutional sub-models at word and at sentence level that are end to end trainable. We show that state-of-the-art (SOTA) Transformer models including pretrained BERT do not necessarily yield the best performance for Bengali and Hindi. CoCNN outperforms pretrained BERT with 16X less parameters and achieves much better performance than SOTA LSTMs on multiple real-world datasets. This is the first study on the effectiveness of different architectures from Convolution, Recurrent, and Transformer neural net paradigm for modeling Bengali and Hindi.

Název v anglickém jazyce

CNN for Modeling Sanskrit Originated Bengali and Hindi Language

Popis výsledku anglicky

Though recent works have focused on modeling high resource languages, the area is still unexplored for low resource languages like Bengali and Hindi. We propose an end to end trainable memory efficient CNN architecture named CoCNN to handle specific characteristics such as high inflection, morphological richness, flexible word order and phonetical spelling errors of Bengali and Hindi. In particular, we introduce two learnable convolutional sub-models at word and at sentence level that are end to end trainable. We show that state-of-the-art (SOTA) Transformer models including pretrained BERT do not necessarily yield the best performance for Bengali and Hindi. CoCNN outperforms pretrained BERT with 16X less parameters and achieves much better performance than SOTA LSTMs on multiple real-world datasets. This is the first study on the effectiveness of different architectures from Convolution, Recurrent, and Transformer neural net paradigm for modeling Bengali and Hindi.

Druh

D - Stať ve sborníku

CEP obor

—

OECD FORD obor

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

Projekt

—

Návaznosti

—

Rok uplatnění

2022

Kód důvěrnosti údajů

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

Název statě ve sborníku

Proceedings of the 2nd Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics and the 12th International Joint Conference on Natural Language Processing (Volume 1: Long Papers)

ISBN

978-1-955917-65-0

ISSN

—

e-ISSN

—

Počet stran výsledku

10

Strana od-do

47-56

Název nakladatele

Association for Computational Linguistics

Místo vydání

—

Místo konání akce

Online only

Datum konání akce

1. 1. 2022

Typ akce podle státní příslušnosti

WRD - Celosvětová akce

Kód UT WoS článku

—