{"id":1144,"date":"2024-02-12T14:03:50","date_gmt":"2024-02-12T13:03:50","guid":{"rendered":"https:\/\/www.cjvt.si\/povejmo\/?page_id=1144"},"modified":"2024-07-05T10:23:18","modified_gmt":"2024-07-05T08:23:18","slug":"work-package-4-vemo-med","status":"publish","type":"page","link":"https:\/\/www.cjvt.si\/povejmo\/en\/work-packages\/work-package-4-vemo-med\/","title":{"rendered":"Work Package 4: VeMo-Med"},"content":{"rendered":"

Adaptation of Large Language Models for Use in Slovenian Medicine<\/h2>\n<\/div><\/section>
\n

VeMo-Med<\/h3>\n<\/div><\/section><\/p><\/div>

In this project, we will extend the use of language models to the field of medicine. Better d.o.o. offers electronic prescription options, automatic speech transcriptions for discharge letters, identification of active ingredients\/medications, dosages, time constraints, and stakeholders (doctors, patients, others); automatic digitization of forms with text field and input recognition.<\/p>\n<\/div><\/section><\/div><\/p>\n

Specific objectives:<\/h3>\n
    \n
  1. Creation of command and dialogue-based training datasets with specific dialogs and commands related to medical applications.<\/li>\n
  2. Development of a large generative Slovenian language model tailored for medicine.<\/li>\n
  3. Creation of a precise Slovenian speech recognizer specialized for the medical field.<\/li>\n
  4. Development of an advanced medical application utilizing speech recognition and large language models for more efficient and quicker healthcare personnel tasks.<\/li>\n<\/ol>\n<\/div><\/section><\/div>\n

    Results:<\/h3>\n