Leveraging Large Language Models for Korean Sign Language Translation

Abstract

In the field of sign language translation, the use of neural network architectures has been identified as a crucial factor in improving performance. Large Language Models (LLMs) are particularly promising in this regard. This study investigates the potential of using LLMs for sign language translation and examines various training methodologies to identify key strategies. The use of extensive linguistic data in LLMs has been shown to improve translation accuracy. When compared to traditional translation models, the benefits of the LLM-based approach are clear. LLMs are particularly effective at capturing subtle nuances of sign language. Future research will explore more nuanced applications of these models. These findings contribute to the ongoing discussion on optimizing sign language translation methods.