Explainability

Explore cutting-edge research on explainable AI in this 50-minute conference session from the ACM FAT* 2019 conference. Chaired by Giles Hooker, the session features four presentations covering diverse aspects of AI explainability: actionable recourse in linear classification, model reconstruction from explanations, efficient search for diverse coherent explanations, and a case study on human predictions with explanations versus machine learning models in deception detection. Gain insights into the latest advancements and challenges in making AI systems more transparent and interpretable.