Leveraging NLP and Neuro-Symbolic AI for Early Diagnosis and Causal Inference in Mental Health Disorders

Samarth Yogesh Jadhav, Rutuja Rajaram More

Keywords: Mental Health, Artificial Intelligence, Natural Language Processing, Sentiment Analysis, Emotion Detection, Machine Learning

TL;DR: We propose a framework combining Natural Language Processing and Neuro-Symbolic AI to enable early diagnosis and interpretable causal inference for mental health disorders using unstructured text data.

Abstract: Mental illnesses, such as depression, anxiety, and schizophrenia, rank among the first three causes of disability around the globe. Though critical to treatment, timely and accurate diagnosis is seldom afforded by current systems because of their time-consuming nature, subjectivity, and reliance on clinical expertise. With this overview, we suggest an intersection between Natural Language Processing (NLP) and Neuro-Symbolic AI to further advances in mental health diagnostics and causal inference. It focuses on the technology enabling NLP to analyze unstructured texts such as social media postings, transcripts from therapy, and clinical records that are considered indicators of mental conditions. Neuro-symbolic AI may further remediate this provided it could offer interpretable, causality-relevant models to intuit the onset of conditions related to mental health. In conclusion, we introduce an integrative framework that marries the strengths of these technologies to further diagnostics and causal understanding.

Submission Number: 1