Multi-Lingual Representation of Natural Language Processing for Low Resource Asian Language Processing Systems
Elena Verdú, Yuri Vanessa Nieto, Nasir Saleem · ACM Transactions on Asian and Low-Resource Language Information Processing · 2024
Natural Language Processing (NLP) has undergone a significant shift towards hybrid approaches that combine probabilistic and symbolic methods in recent years.However, building an efficient NLP model for low-resource languages can be challenging, particularly for Asian languages with various dialects and limited resources.This challenge can be addressed with multilingual representation techniques that allow for knowledge transfer from adequate training data to low-resource languages.These techniques can simplify language processing tasks and reduce complexity and cost measures.The use of transfer learning and neural networks for multi-lingual language processing can help understand semantic similarities and enable effective communication between machines and users in various languages.Therefore, it is crucial to focus on multi-lingual representation for NLP models that contain low-resource Asian languages to scale effectively.This special issue aims to address this challenge by exploring multi-lingual representation techniques for NLP in low-resource Asian languages.In this special issue four papers were accepted for publication after a rigorous review process.The following are the highlights of the noteworthy technological discoveries of the accepted works.In the first paper titled "More than Syntaxes: Investigating Semantics to Zero-shot Cross-lingual Relation Extraction and Event Argument Role Labelling" the authors introduce Syntax and Semantic Driven Network (SSDN) to equip syntax and semantic knowledge across languages simultaneously.Specifically, predicate-argument structures from semantic role labelling are explicitly incorporated into word representations.The evaluation results demonstrate that the proposed method achieves the state-of-the-art performance.Further study also indicates SSDN could produce robust representations that facilitate the transfer operations across languages.In the second paper titled "A Research on University Students' Behavioral Intention to Use New- generation Information Technology in Intelligent Foreign Language Learning", in order to investigate the factors influencing university students using a new generation of information technology in intelligent foreign language learning, the author proposes a research model based on Technology Acceptance Model (TAM).The sample data were collected from a survey of 237 students at a university.The analysis of structural equation modeling indicated that perceived usefulness, perceived ease of use, construction of foreign language intelligence classroom, and computer selfefficacy all have a positive impact on learners' behavioral intention.These findings suggest that more emphasis should be laid on students' learning characteristics, advantages of a new generation of technology, as well as teachers' intelligent literacy, so as to further promote the deep integration