Causality-based Explanation of Classification Outcomes
arXiv:2003.06868v249 citations
AI Analysis
This work addresses the need for interpretable AI in financial applications, but appears incremental as it builds on prior explanation methods.
The authors tackled the problem of explaining classifier outcomes by proposing a causality-based definition, comparing it with existing notions and analyzing their complexity, and conducted experiments on two financial datasets.
We propose a simple definition of an explanation for the outcome of a classifier based on concepts from causality. We compare it with previously proposed notions of explanation, and study their complexity. We conduct an experimental evaluation with two real datasets from the financial domain.