Neuro-symbolic AI for mental health and well-being

Abstract

Neuro-Symbolic AI (NSAI)- the amalgamation of Neural AI with Symbolic AI leverage's the approach of traditional machine learning models where the system of NSAI gives output by learning data’s behavior using neural network learning techniques and symbolic reasoning converts the behaviors into human understandable symbols so that the reasoning is transparent. With the advancement of technology, mental illness is one of the inevitable issues that has already suppressed the magnitude of physical sickness. Considering the reputation of artificial intelligence, traditional machine learning techniques are the first contenders for the early detection of mental disorders despite their limitations such as lack of explainability, poor performance on imbalanced data, and a small amount of data. The limitations mislead the results for diagnosing and assisting mental diseases. In addition, Neuro-symbolic AI can be the holistic approach to prevent the lack of precision in existing machine learning models with small amounts of data and detect mental health patients by giving proper explanations as a handcrafted expert does. Our proposed model of Neuro-symbolic AI predicts Alzheimer’s disease with 95.45% accuracy which decisively beats the result of traditional machine learning models. Therefore, the paper aims to find a better approach to the concerned people and play the role of assistant to detect the disease earlier and make their lives safe.