Disease Diagnosis On Ships Using Hierarchical Reinforcement Learning | IEEE Conference Publication | IEEE Xplore

Disease Diagnosis On Ships Using Hierarchical Reinforcement Learning


Abstract:

Every year about 30 million people travel by ship worldwide often in extreme weather conditions and polluted environments and many other factors that impact the health of...Show More

Abstract:

Every year about 30 million people travel by ship worldwide often in extreme weather conditions and polluted environments and many other factors that impact the health of passengers and crew staff. Such issues require medical staff for passenger health care. We introduce a model based on Reinforcement learning(RL) which is used in the dialogue system. We incorporate the Hierarchical reinforcement learning (HRL) model with the layers of Deep Q-Network for dialogue oriented diagnosis system. Policy learning is integrated as policy gradients are already defined. We created a two-stage hierarchical strategy. We used the hierarchical structure with double-layer policies for automatic disease diagnosis. A double layer means it splits the task into sub-tasks named high-state strategy and low-level strategy. It has a user simulator component that communicates with the patient for symptom collection low-level agents inquire about symptoms. Once it’s done collecting it sends results to the high-level agent which activates the D-classifier for the last diagnosis. When it’s done its sent back by the user simulator to patients to verify the diagnosis made. Every single diagnosis made has its reward that trains the system
Date of Conference: 08-11 September 2024
Date Added to IEEE Xplore: 04 November 2024
ISBN Information:
Conference Location: Belgrade, Serbia

References

References is not available for this document.