Skip to main content

Introduction to Explainable AI Research

Our research laboratory focuses on developing innovative approaches to Explainable Artificial Intelligence (XAI) that bridge the gap between powerful AI models and human understanding. As AI systems become increasingly integrated into critical decision-making processes across various domains, the need for transparency, interpretability, and trustworthiness has never been more essential.

Our work spans four interconnected research themes that address different facets of the explainability challenge:

  1. Explainable Predictive Process Analytics: We develop methods to reveal the reasoning behind business process predictions, enhancing trust and providing actionable insights for organizations. Our approaches enable interpretability throughout the predictive analytics pipeline, from feature construction to model inspection and user-centric explanations.

  2. Probabilistic & Causal Models for Responsible AI: We focus on making complex statistical models understandable through causal reasoning and counterfactual explanations. By incorporating "what-if" scenarios, we enable users to understand not just correlations but the causal mechanisms behind AI decisions.

  3. Persuasive Models for Explainable AI: We transform technical AI outputs into compelling, trustworthy information through narrative-based explanations. Our research bridges the gap between symbolic information extracted from black-box models and human-understandable explanations that build trust.

  4. Explainable Medical Diagnostic Systems: We build human-centric explainable interfaces that assist physicians in understanding AI predictions for medical image diagnosis. Our frameworks support the generation of intuitive explanations that translate complex model outputs into actionable clinical insights.

These research directions share a common goal: to develop AI systems that not only make accurate predictions but can also communicate their reasoning in ways that are meaningful, trustworthy, and actionable for human users. Through this work, we aim to advance the field of XAI and contribute to the responsible development and deployment of AI across healthcare, industry, government, and other critical sectors.