top of page
JONAS
A Vaccine Concerns Assessment Tool
Technology
Architecture
As the patient chats with Jonas, responses are logged to Google BigQuery via a Firebase cloud function. The model is hosted in Django, which will evaluate the conversation results. Finally, model results will be logged back into Google BigQuery. All model results are then displayed into our Salesforce Vaccine Hesitancy Dashboard for medical provider viewing and verification.
Model
Trained on vaccine-related tweets from Twitter and self-collected surveys via Qualtrics, Jonas employs XLNet (1). XLNet assists Jonas in detecting up to 9 vaccine hesitancy concerns from only five simple survey questions.
(1) For more information on XLNet, please view the original paper on arXiv.
bottom of page