A secured self-learning counselling system leveraging RNN

Abstract

Chatbots are virtual assistants providing round-the-clock service across many sec tors globally. It helps manage human workforce scarcity. As such, the need for mental health care services has been on the rise. Currently, the demand exceeds the availability of such services. Moreover, in many parts of the world, mental health care is still inaccessible due to various socio-economic reasons. In this research, we propose a Chatbot to aid in providing mental health care services that is accessible from anywhere in the world. Our model will be capable of providing accurate psy chological health care as well as ensuring the privacy of data that users share with the Chatbot. We have used RNN to train the model of this Chatbot. Many users’ conversations can be used to train this Chatbot easily as we have followed the RNN algorithm in a way that allows us to train large quantity of data in a comparatively less powerful machine. However, it is imperative to stress that these conversations are solely used for training the model. No one except users will be aware of the conversations with the Chatbot. Moreover, we plan to implement self-learning tech nology to combat the insufficiency of data necessary for training our model. The more it is used, the better it will get, learning from the user’s responses.