The research focuses on Explanatory Model Steering (EXMOS), a method to fine-tune prediction models using Explainable AI and Interactive Machine Learning. The study investigates the effectiveness of global model-centric and data-centric explanations in aiding domain experts, particularly healthcare experts, to detect and resolve potential data issues for model improvement. The results illustrate that a hybrid fusion of both explanation types demonstrates the highest effectiveness. The study also provides design implications for effective explanation-driven interactive machine-learning systems.
Publication date: 1 Feb 2024
Project Page: https://doi.org/10.1145/3613904.3642106
Paper: https://arxiv.org/pdf/2402.00491
