همایش بین المللی هوش مصنوعی

صفحه اصلی / اولین همایش بین المللی هوش مصنوعی

Persian Intelligent Assistant in Healthcare Domain

نویسندگان :

Sarina Chitsaz¹ Mehrnoush Shamsfard²

1- دانشگاه شهید بهشتی 2- دانشگاه شهید بهشتی

کلمات کلیدی :

Intelligent Assistant،Natural language understanding،Speech act classification،Slot filling

چکیده :

Nowadays, advances in technology and medical science have led to significant changes in the field of healthcare. One of the best examples of these developments is the emergence of intelligent assistants in this field. These assistants, powered by natural language processing and AI, utilize medical data to efficiently organize and deliver crucial patient information. Particularly in emergency departments and triage units, these assistants can enhance the accuracy and speed of diagnosing and prioritizing patients' health conditions, making them vital tools in emergency healthcare management. In this research, an attempt has been taken to provide a framework for designing an intelligent health assistant in Persian language. Therefore, the first step was to prepare a labeled dataset related to the mentioned domain. In the next step, the intelligent assistant architecture was designed and implemented. The natural language understanding component of the assistant is based on slot filling and speech act classification. Moreover, the dialogue manager section was created to address the challenge of negation in patients' utterances. In the end, the output of the previous steps is the classification of the triage patients. By the time, the assistant's performance was evaluated, the results showed consistency with the performance of the emergency staff in 83% of cases.