A Corpus-Based Auto-encoder-and-Decoder Machine Translation Using Deep Neural Network for Translation from English to Telugu Language

• Published in: SN computer science Vol. 4; no. 4; p. 354
• Main Authors: Mahanty, Mohan, Vamsi, Bandi, Madhavi, Dasari
• Format: Journal Article
• Language: English
• Published: Singapore Springer Nature Singapore 01.07.2023; Springer Nature B.V
• Subjects: Artificial neural networks; Coders; Computer Imaging; Computer Science; Computer Systems Organization and Communication Networks; Data analysis; Data processing; Data Structures and Information Theory; Datasets; Enabling Innovative Computational Intelligence Technologies for IOT; English language; Information Systems and Communication Service; Language; Linguistics; Machine learning; Machine translation; Multilingual systems; Multilingualism; Neural networks; Original Research; Pattern Recognition and Graphics; Sentences; Software Engineering/Programming and Operating Systems; Syntax; Translating; Translations; Vision; India; Deep neural network; English-to-Telugu; Neural machine translation; Encoder-and-decoder
• ISSN: 2661-8907; 2662-995X; 2661-8907
• DOI: 10.1007/s42979-023-01678-4

There is a huge demand for machine translation to design automated auto-encoders that can convert English to the Telugu language. Neural machine translation (NMT) is a new and effective technology that has significantly benefited traditional machine translation strategies. As a result, the primary goal of this work is to propose an auto-encoder–decoder-based deep neural network for translating English to the Telugu language. Deep neural networks are used for learning, where input language is supplied to the neural network following data processing and analysis. The neural network then does auto-tuning, which helps to improve the translation quality. This proposed work produces target sentences in the Telugu language, resulting in a higher Bilingual Evaluation Understudy (BLEU) Score and a lower Word Error Rate (WER). The BLEU finds the similarity of the translated text by Machine Translation models with high-quality references and WER finds better accuracy in identifying the target sentence.