Speech-to-speech Low-resource Translation

Speech-to-speech translation (S2ST), particularly in the context of low-resource languages, plays a vital role in facilitating global communication. However, comprehensive research in this emerging field is lacking, especially concerning translation without the use of text. The objective of this study is to bridge the gap by conducting a systematic review of existing literature on S2ST for low-resource languages. We discovered 455 articles by searching the Scopus, IEEE Xplore, and ACM Digital Library databases, focusing on identifying research trends. The results highlight significant topics covered in the literature, marking a transition from traditional neural network methodologies to advanced transformer-based models. Our findings provide a robust overview of the S2ST landscape, identifying challenges and potential solutions for future research, particularly regarding the application of this technology in low-resource settings. The research contribution of this study is the insights gleaned will benefit academics and professionals seeking a comprehensive understanding of S2ST for low-resource languages.