Scholarlink

<p>The accurate evaluation and efficient treatment of mental health problems, such as depression, continue to be difficult due to the complicated underlying causes and diverse signs and symptoms. In this study, the integration of Large Language Models (LLMs) and Explainable AI (XAI) offers a transformative approach to mental health care, surpassing traditional systems in conversational relevance, empathy, and risk detection. The proposed hybrid framework leverages LLMs’ contextual understanding and XAI’s transparency to deliver inclusive, personalized, and trustworthy mental health interventions. BioBERT (bio-medical bidirectional encoder representations from transformers) is being used as LLM which produces vector representations for questions and answers, encapsulating the semantic content and context of biomedical language. These vector representations are then saved for future use, guaranteeing efficient retrieval during the query phase. It is followed by utilization of bi-directional gated recurrent unit (Bi-GRU) to analyze vector representations in both forward and backward directions, therefore capturing temporal connections essential for comprehending emotions and contextual refinements.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Scalable and transparent mental health support via XAI-LLM

  • 期刊论文

  • 发表日期：

    2025年6月21日

查看全文
  • K. Nimala,
  • Chin-Shiuh Shieh,
  • R. Nareshkumar,
  • V. Senthil Murugan

摘要

The accurate evaluation and efficient treatment of mental health problems, such as depression, continue to be difficult due to the complicated underlying causes and diverse signs and symptoms. In this study, the integration of Large Language Models (LLMs) and Explainable AI (XAI) offers a transformative approach to mental health care, surpassing traditional systems in conversational relevance, empathy, and risk detection. The proposed hybrid framework leverages LLMs’ contextual understanding and XAI’s transparency to deliver inclusive, personalized, and trustworthy mental health interventions. BioBERT (bio-medical bidirectional encoder representations from transformers) is being used as LLM which produces vector representations for questions and answers, encapsulating the semantic content and context of biomedical language. These vector representations are then saved for future use, guaranteeing efficient retrieval during the query phase. It is followed by utilization of bi-directional gated recurrent unit (Bi-GRU) to analyze vector representations in both forward and backward directions, therefore capturing temporal connections essential for comprehending emotions and contextual refinements.