AI-Based Multilingual Machine Translator from English to GorBoli
AI-Based Multilingual Machine Translator from English to GorBoli
Authors:
Ch. Poojitha Department of CSE RGUKT Basar, India
- Poojitha Department of CSE RGUKT Basar, India
- Pinki Department of CSE RGUKT Basar, India
Mrs. P. Sarika Rao Assistant Professor RGUKT Basar, India
Abstract—In the era of rapid digital transformation, multilingual communication has become a fundamental requirement for enabling seamless interaction across di- verse linguistic communities. Although significant progress has been made in Natural Language Processing (NLP) and Artificial Intelligence (AI), most existing machine translation systems predominantly focus on high-resource languages such as English, Hindi, and Chinese. This creates a substantial digital divide for speakers of low-resource regional languages such as GorBoli, which lack sufficient datasets, computational resources, and pretrained models. This paper presents an AI-based multilingual machine
translation system designed to translate English text into GorBoli using a hybrid approach. The proposed system integrates transformer-based deep learning models, specifically the T5 architecture, along with external translation APIs to support multiple languages. In addition to text- based translation, the system incorporates speech recognition and text-to-speech technologies, enabling users to interact through both voice and text interfaces.
Due to the absence of publicly available datasets for Gor- Boli, a custom dataset consisting of English-GorBoli sentence pairs was created and preprocessed using techniques such as tokenization, normalization, and noise removal. The system architecture follows a pipeline-based design, including input acquisition, preprocessing, translation, and output generation. The backend is implemented using Flask, with RESTful APIs handling chatbot interaction, translation, and speech processing.
Experimental evaluation demonstrates that the pro- posed system achieves an accuracy of approximately 90% for simple and moderately complex sentences, with efficient response time and improved usability. The integration of multilingual and speech-based capabilities enhances accessibility, particularly for users unfamiliar with typing or standard languages.
Overall, this work contributes toward bridging the language gap by providing an effective, scalable, and user- friendly solution for low-resource language translation. It highlights the importance of inclusive AI systems and opens opportunities for future research in expanding datasets, improving contextual understanding, and integrating advanced large language models for enhanced performance.
Index Terms—Machine Translation, NLP, Transformer Models, Low-Resource Language