A Survey of Mental Health Chatbots using NLP

Abstract

Due to recent technological advancements, there has been a rise in digital interventions aimed at either supplementing or replacing face-to-face mental health services which include the advent of mental health chatbots, that claim to provide assistive care using a therapeutic approach. These innovations are driven by Natural Language Processing (NLP). A chatbot can be described as a computer program capable of providing smart responses to user inputs by understanding natural language using one or more NLP techniques. In this study, we discuss the application of NLP in psychotherapy and also a general analysis of existing systems was performed by comparing the responses given by the chatbot against a set of predefined user inputs pertaining to queries related to wellbeing and mental health. The general methodology involved in the creation of such chatbots includes the underlying NLP techniques like word embeddings, sentiment analysis, Models like Sequence-to-Sequence model and attention mechanism. We also propose MentalEase, a mobile application which uses NLP techniques to provide not only conversational aid but also a toolbox of helpful features to keep mental health in place. By integrating mental health assessment tools into the chatbot interface, along with regular therapy it can help patients deal with mild anxiety and depression. This can also overcome certain barriers in mental health, such as waiting lists and geographical problems that hinder attendance at face-to-face counselling appointments.