Translating Sanskrit to Hindi Language using Recurrent Neural Network (RNN)-L2 Regularization
Year: 2024
Type: article
Abstract: Machine Translation (MT) is a subfield of computer linguistics that focuses on the automatic translation from one natural language into another without any human involvement. There is a huge need for translating information between languages to send and communicate thoughts because native people interact in a variety of languages. However, Sanskrit is an ancient Indo-European language that requires essential processing to be explored in computer science and computational language analysis. In this paper, Recurrent Neural Network (RNN)-L2 regularization method is proposed for the Sanskrit to Hindi translation language. A neural machine translation system is trained using the linguistic information from the rule-based input. The proposed method is innovative and adaptable to any low-resource language with extensive morphology that covers multiple domains with minimal human involvement. The efficacy of the RNN-L2 regularization method is demonstrated by employing the dataset of Corpora. The existing methods such as machine translation systems, and hybrid machine translation systems are used to explain the efficacy of the RNN-L2 regularization method. The proposed RNN-L2 regularization method achieves better BLEU score, and METEOR of 76% and 72% compared with the existing methods such as machine translation systems, and hybrid machine translation systems.
Institution: Visvesvaraya Technological University
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W7217047129