Bridging the Gap: Advancing Specialized Translation with Neural Machine Translation in Technical Fields – A Practical Approach for Small Businesses.

This thesis examines the intersection of Language for Specific Purposes (LSP), specialized translation, and the advancements in Machine Translation (MT), with a focus on Neural Machine Translation (NMT) and its practical implementation in day-to-day activities in small businesses. Chapter One explores the need for specialized terminology in technical and scientific fields, highlighting the challenges and relevance of specialized translation. It addresses the role of standardization in ensuring accuracy and consistency across different domains. Chapter Two provides a historical overview of MT, tracing its evolution from rule-based approaches to the emergence of NMT. It dissects key components of NMT, such as the encoder-decoder architecture and the transformer model, which have significantly enhanced translation quality. The chapter also addresses the challenges of applying NMT to multilingual contexts and low-resource languages, offering solutions for these issues. Chapter Three shifts to the practical application of NMT within specialized technical fields, particularly in small companies. It outlines the workflow of using NMT, including the collection of relevant materials, pre-editing, the NMT process, and post-editing. The chapter emphasizes the importance of quality assurance, using automatic evaluation metrics, and considering language pair coverage and domain specialization. Real-world examples from the electrical and electronic sectors demonstrate how NMT was employed to translate highly technical texts, showcasing its effectiveness and the challenges encountered. Overall, this thesis provides a comprehensive analysis of the relationship between specialized translation and MT, offering both theoretical insights and practical applications.