In our quest to provide the best services to our clients and utilize the abilities of machine translation to the utmost, we are always in the process of R&D, working, in particular, with colleagues from the University of Manchester. For two years now, we focused our efforts on medical text machine translation, multilingual neural network models, and experiments with transfer learning methodology. 

We worked with both the massive multilingual pre-trained language models (MMPLMs) and our proprietary small-sized pre-trained language model (PLM) – and achieved top-level results on translation of clinical domain data in the English-Spanish language pair. Moreover, we have demonstrated that a small-sized PLM shows much better results than the extra-large language models. 

All of these findings were summarized in the research paper that was submitted for publication – and we are happy to announce that the paper has been accepted by the Frontiers in Digital Health magazine[1]! This is great news for our R&D team and our colleagues at the University of Manchester, and an acknowledgement that we are moving in the right direction. Stay tuned for more!

__________________________________________________________________________________

[1] “Frontiers in Digital Health” is a peer-reviewed journal at the forefront of health and technology. It covers a wide range of topics including connected health, digital mental health, health informatics, and personalized medicine, among others. Following an open access model, this journal ensures rigorous peer review for maintaining high scientific standards, providing a vital platform for researchers in digital health. For more information, visit their website.