Machine Transliteration

Machine transliteration focuses on automatically converting text from one writing system to another while preserving pronunciation, a crucial task for bridging language barriers in multilingual applications. Current research emphasizes improving cross-lingual alignment in multilingual language models (MLLMs) using transliteration, often employing transformer-based architectures and contrastive learning techniques to enhance representation similarity between different scripts. This work is significant for advancing cross-lingual understanding in NLP, particularly for low-resource languages and those with diverse scripts, and has practical implications for improved machine translation, speech recognition, and text processing across diverse linguistic contexts.