A Chatbot for the Staff in Radiotherapy Using Artificial Intelligence and Machine Learning
K Li1, J Chow2*, (1) York University, Toronto , ON, CA, (2) Princess Margaret Cancer Centre, Toronto, ON, CA.
Presentations
PO-GePV-E-1 (Sunday, 7/25/2021) [Eastern Time (GMT-4)]
Purpose: A chatbot personalized by artificial intelligence and machine learning is built to provide educational information for staff in radiotherapy. The topics covered include interdisciplinary knowledge in radiation oncology, medical physics, radiotherapy, and radiation safety.
Methods: The Bot was constructed using the IBM Watson Assistant functionalities on the IBM cloud. The Bot had ability to communicate with radiation staff by answering some simple questions regarding radiotherapy, or by multiple choice to the user from a list of options. A layered structure approach was used in the workflow of the Bot to interact with the user. Through selecting different topics in the list and communication between the Bot and user, he/she will finally acquire the requested information. In addition, communications between the Bot and user will be used in machine training leading to a more accurate response to the user’s question.
Results: The user interface of the Bot is a front-end window operating on an internet of things such as smartphone, tablet or laptop. The user can type in a question or anything to communicate with the Bot. In the communication, if the Bot cannot identify what the user needs, the Bot will provide a list of options to the user as guidance. Through selecting different topics on the list, another list with more detailed breakdown of the topic will pop up, so that the user can obtain his/her request accurately and quickly. The Bot acquires knowledge through interactions with the users.
Conclusion: It is concluded that a chatbot with characterization supported by machine learning can provide information regarding radiotherapy to the radiation staff more efficiency, and can be used in staff training. This Bot is particularly useful for quick search such as looking for physics data or a review of some radiation safety protocol conveniently.
Funding Support, Disclosures, and Conflict of Interest: This work is currently supported by the Planning and Dissemination Grants, Institute Community Support, Canadian Institute of Health Research.
ePosters
Keywords
Radiation Therapy, Radiation Protection
Taxonomy
Not Applicable / None Entered.
Contact Email