AI-Powered Virtual Mental Health Assistant for Early-Stage NLP-Based Mental Health Screening

Desmond Yong Kiat Loy, Peter Chunyu Yau, Dennis Wong

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Mental health issues have become a significant global concern, affecting individuals' ability to perform daily tasks and increasing the risk of social prejudice. This paper presents a solution to address mental health challenges while reducing the burden on medical professionals. Our research includes a survey of 41 respondents to gauge public perception of mesntal health and their willingness to share their thoughts. We propose a web-based platform that provides mental health guidance and facilitates direct communication with psychologists. The platform integrates a speech-to-text model for audio transcription and a natural language processing (NLP) model to classify mental health conditions. Its architecture ensures secure data storage while enabling users to access essential resources without fear of discrimination.

Original languageEnglish
Title of host publication2025 3rd Cognitive Models and Artificial Intelligence Conference, AICCONF 2025 - Proceedings
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9798331509699
DOIs
Publication statusPublished - 2025
Event3rd Cognitive Models and Artificial Intelligence Conference, AICCONF 2025 - Prague, Czech Republic
Duration: 13 Jun 202514 Jun 2025

Publication series

Name2025 3rd Cognitive Models and Artificial Intelligence Conference, AICCONF 2025 - Proceedings

Conference

Conference3rd Cognitive Models and Artificial Intelligence Conference, AICCONF 2025
Country/TerritoryCzech Republic
CityPrague
Period13/06/2514/06/25

Keywords

  • Mental Health
  • Natural Language Processing (NLP)
  • RoBERTa
  • Video Streaming
  • WebRTC

Fingerprint

Dive into the research topics of 'AI-Powered Virtual Mental Health Assistant for Early-Stage NLP-Based Mental Health Screening'. Together they form a unique fingerprint.

Cite this