Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

AXAI-CDSS : An Affective Explainable AI-Driven Clinical Decision Support System for Cannabis Use

Tongze Zhang, Tammy Chung, Anind Dey, Sang Won Bae

TL;DR

This work tackles the need for transparent, affective AI-driven clinical decision support for cannabis use by integrating explainable AI (XAI), causal inference, and large language model (LLM) explanations with real-time emotion sensing. It employs per-user multimodal models trained on smartphone and wearable data, presenting explanations through SHAP, causal diagrams, counterfactuals, and SkopeRules, with integrated dialogue via GPT-4. The system adapts its tone and content based on clinicians' emotions detected from facial expressions and sentiment analysis to reduce cognitive load and build trust. An evaluation demonstrates improved usability, personalization, and clarity for the emotion-aware, prompt-engineered version over a baseline, highlighting practical benefits and challenges for deployment in healthcare. The study emphasizes trustworthy, human-centered AI in clinical settings while acknowledging privacy, connectivity, and computational considerations for real-world implementation.

Abstract

As cannabis use has increased in recent years, researchers have come to rely on sophisticated machine learning models to predict cannabis use behavior and its impact on health. However, many artificial intelligence (AI) models lack transparency and interpretability due to their opaque nature, limiting their trust and adoption in real-world medical applications, such as clinical decision support systems (CDSS). To address this issue, this paper enhances algorithm explainability underlying CDSS by integrating multiple Explainable Artificial Intelligence (XAI) methods and applying causal inference techniques to clarify the model' predictive decisions under various scenarios. By providing deeper interpretability of the XAI outputs using Large Language Models (LLMs), we provide users with more personalized and accessible insights to overcome the challenges posed by AI's "black box" nature. Our system dynamically adjusts feedback based on user queries and emotional states, combining text-based sentiment analysis with real-time facial emotion recognition to ensure responses are empathetic, context-adaptive, and user-centered. This approach bridges the gap between the learning demands of interpretability and the need for intuitive understanding, enabling non-technical users such as clinicians and clinical researchers to interact effectively with AI models.} Ultimately, this approach improves usability, enhances perceived trustworthiness, and increases the impact of CDSS in healthcare applications.

AXAI-CDSS : An Affective Explainable AI-Driven Clinical Decision Support System for Cannabis Use

TL;DR

This work tackles the need for transparent, affective AI-driven clinical decision support for cannabis use by integrating explainable AI (XAI), causal inference, and large language model (LLM) explanations with real-time emotion sensing. It employs per-user multimodal models trained on smartphone and wearable data, presenting explanations through SHAP, causal diagrams, counterfactuals, and SkopeRules, with integrated dialogue via GPT-4. The system adapts its tone and content based on clinicians' emotions detected from facial expressions and sentiment analysis to reduce cognitive load and build trust. An evaluation demonstrates improved usability, personalization, and clarity for the emotion-aware, prompt-engineered version over a baseline, highlighting practical benefits and challenges for deployment in healthcare. The study emphasizes trustworthy, human-centered AI in clinical settings while acknowledging privacy, connectivity, and computational considerations for real-world implementation.

Abstract

As cannabis use has increased in recent years, researchers have come to rely on sophisticated machine learning models to predict cannabis use behavior and its impact on health. However, many artificial intelligence (AI) models lack transparency and interpretability due to their opaque nature, limiting their trust and adoption in real-world medical applications, such as clinical decision support systems (CDSS). To address this issue, this paper enhances algorithm explainability underlying CDSS by integrating multiple Explainable Artificial Intelligence (XAI) methods and applying causal inference techniques to clarify the model' predictive decisions under various scenarios. By providing deeper interpretability of the XAI outputs using Large Language Models (LLMs), we provide users with more personalized and accessible insights to overcome the challenges posed by AI's "black box" nature. Our system dynamically adjusts feedback based on user queries and emotional states, combining text-based sentiment analysis with real-time facial emotion recognition to ensure responses are empathetic, context-adaptive, and user-centered. This approach bridges the gap between the learning demands of interpretability and the need for intuitive understanding, enabling non-technical users such as clinicians and clinical researchers to interact effectively with AI models.} Ultimately, this approach improves usability, enhances perceived trustworthiness, and increases the impact of CDSS in healthcare applications.

Paper Structure

This paper contains 17 sections, 4 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Advantages of Language Models in designing CDSS-HCI
  4. XAI in CDSS-HCI
  5. Prompt Engineering for AI Interactions
  6. Facial emotion recognition in CDSS
  7. Gaps in health risk monitoring, prediction and intervention
  8. Method
  9. Data Collection
  10. Model Generation
  11. Interpretability Analysis
  12. Prompt Engineering
  13. Dialogue System
  14. Result
  15. Discussion
  16. ...and 2 more sections

Figures (4)

  • Figure 1: Overview of the Affective Explainable AI-Driven Clinical Decision Support System (AXAI-CDSS)
  • Figure 2: Comparison of Our Proposed Real-Time Emotion-Aware Conversational AI (Left) and Basic LLMs (Right)
  • Figure 3: System Demonstration (Top) and Example of Multi-AI Explainability with Causal Reasoning (Bottom)
  • Figure 4: Comparative Results across Different Metrics