Statistical Machine Translation for Multilingual Access to Medical AI Education Content

Statistical Machine Translation (SMT), as part of an overall human-centered machine translation process, provides a significant opportunity for multilingual access to digital content, such as specialized educational content in medical artificial intelligence. As medical knowledge is a continually evolving field of study and is critical for timely and accurate access by global learners and health professionals, providing accurate and accessible learning materials across many languages makes sense. The traditional translation solutions including online machine translation utilize outdated methods of translation that do not account for specific or context-relevant solutions, and domain-specific languages, terminologies, and linguistic affordances or preferences. Rule-based machine and general-purpose machine translation solutions will be limited in reflecting medical terminology and will ultimately adversely affect the translation quality and educational access provided. Therefore, here suggested a new framework called MED-SMT (Medical Educational Domain Statistical Machine Translation) to deal with these problems. MED-SMT is a domain adapted statistical translation model, with an established medical terminology database and a set of contextually grounded alignment processes that will translate medical common language. MED-SMT supports the translation of various medical specific language patterns and provides accurate multi-linguistic rendition of the original medical course content. The MED-SMT framework aims to allow the worldwide access to medical Artificial Intelligence (AI) education through the automatic translation of course materials (e.g. lecture notes, research papers, interactive tutorials) into a variety of languages at a minimal cost in human operator time. This will make it easier for medical students, professionals, and researchers who do not speak English to understand advanced AI content. Experimental results show that MED-SMT provides a notable improvement in translation quality and terminology consistency over the advanced methods achieving better understanding and knowledge transfer of medical concepts in medical AI education.

Ҳали таржима қилинмаган