FUSIONLEX: Brigging GAP with Real Time Hybrid Bilingualism
| dc.contributor.author | Ankit Oli 1NH20CE002; Krishnam 1NH20CE024; Pranay 1NH20CE032; Rahul 1NH20CE035 | |
| dc.date.accessioned | 2025-05-31T09:14:47Z | |
| dc.date.available | 2025-05-31T09:14:47Z | |
| dc.date.issued | 2024 | |
| dc.description.abstract | Real-time speech translation has become a transformative technology, enabling instantaneous communication across language barriers. This survey provides an in-depth overview of the current landscape, advancements, challenges, and future directions in real-time speech translation systems. The foundational concepts and methodologies include Automatic Speech Recognition (ASR), which converts spoken language into text: Machine Translation (MT), which translates text from one language to another, and Natural Language Processing (NLP), which facilitates the understanding and processing of human language by computers. These components have evolved significantly, contributing to the progress of real-time speech translation. ASR has transitioned from Hidden Markov Models (HMMs) and Gaussian Mixture Models (GMMs) to deep learning models such as Long Short-Term Memory (LSTM) networks and Convolutional Neural Networks (CNNs), improving accuracy and robustness. MT has shifted from rule-based and statistical models to neural machine translation (NMT) systems, with transformer architectures like BERT and GPT significantly enhancing translation quality. Model optimization and adaptation strategies, such as transfer learning and fine-tuning on specific datasets, enhance performance for particular languages and contexts. Despite significant progress, several challenges remain. Limited training data for many languages hinders the development of accurate translation models. Background noise, accents, and speech variability can affect ASR performance. Ensuring real-time processing without significant delays is a technical challenge, and maintaining high translation accuracy across complex linguistic nuances is difficult. Real-time speech translation is being applied in various fields, including healthcare. business, education, and international diplomacy. In healthcare, it facilitates communication between patients and providers who speak different languages. In business, it enables multinational companies to conduct meetings and negotiations more effectively. In education, it supports learning in multilingual classrooms and access to educational resources in different languages. | |
| dc.identifier.uri | http://192.168.75.5:4000/handle/123456789/19180 | |
| dc.language.iso | en | |
| dc.publisher | NHCE | |
| dc.title | FUSIONLEX: Brigging GAP with Real Time Hybrid Bilingualism | |
| dc.type | Learning Object |