Explainable Artificial Intelligence for Asset Risk Management

Research Objective:

This PhD research focuses on XAI-based proactive asset risk management application for time-series datasets. A group of XAI approaches like SHAP, LIME, and LINDA-BN will be used for an interpretable explanation of the output, and the goal is to provide the risk causes in an interpretable, logical, trustworthy, and optimised way. This research intends to modify LINDA-BN and merge it with a knowledge graph and system dynamics. This merging aims to address and resolve the significant drawback of LINDA-BN, which cannot be applied in time-series datasets. With this update, LINDA- BN can now be used on dynamic, real-time, and complex datasets to provide interpretable output. After AXAI-PRM, this research will design an “Optimised BRB” approach that enforces interpretability in a glass-box manner and offers several advantages over existing XAI approaches. This approach aims to transform a black-box approach into a glass box approach.
 

Achievements:

Journal papers:

S. F. Nimmy, O. K. Hussain, R. K. Chakrabortty, F. K. Hussain, M. Saberi, Explainability in supply chain operational risk management: A systematic literature review, Knowledge-Based Systems 235 (2022) 107587