Data-centric Prediction Explanation via Kernelized Stein Discrepancy
This provides a more effective and robust explanation solution for users needing fine-grained insights into model predictions, though it appears incremental as it builds on existing data-centric approaches.
The paper tackles the problem of coarse-grained and computationally expensive example-based prediction explanations by introducing HD-Explain, a method that uses Kernelized Stein Discrepancy to efficiently identify training samples that best support test predictions, achieving improvements in preciseness, consistency, and efficiency across multiple classification domains.
Existing example-based prediction explanation methods often bridge test and training data points through the model's parameters or latent representations. While these methods offer clues to the causes of model predictions, they often exhibit innate shortcomings, such as incurring significant computational overhead or producing coarse-grained explanations. This paper presents a Highly-precise and Data-centric Explan}ation (HD-Explain) prediction explanation method that exploits properties of Kernelized Stein Discrepancy (KSD). Specifically, the KSD uniquely defines a parameterized kernel function for a trained model that encodes model-dependent data correlation. By leveraging the kernel function, one can identify training samples that provide the best predictive support to a test point efficiently. We conducted thorough analyses and experiments across multiple classification domains, where we show that HD-Explain outperforms existing methods from various aspects, including 1) preciseness (fine-grained explanation), 2) consistency, and 3) computation efficiency, leading to a surprisingly simple, effective, and robust prediction explanation solution.