LAUSR

Dialogue-based simulation is a real-world practice technique for medical and clinical education that provides students with an opportunity to train using hands-on experiences without putting actual patients being put at risk. In this paper, a 3D interactive dialogue-based training and assessment system that supports the detailed development of clinical trial competency for medical students in a distributed virtual environment was proposed. For clinical training, MCRDR-based natural language understanding to realize the semantic representation of written dialog from the most relevant inference results was applied, and on the basis of this, a convolutional neural network model was also used to make the generated inference more exact and reliable. For clinical assessment, the dialogue-driven competency method was used to encompass medical knowledge, communication skill as well as professionalism skill based on the collected dialogue information. Finally, the potential of the created system was demonstrated with several clinical cases. The preliminary results indicate that the system demonstrates the potential of providing efficient training and flexible assessment, while saving time, improving practical skills and making students more confident.