Evaluating Local Model-Agnostic Explanations of Learning to Rank Models with Decision Paths
This work addresses the problem of validating explanation methods for learning-to-rank systems, which is crucial for researchers and practitioners in information retrieval, but it is incremental as it builds on existing explanation techniques.
The paper tackles the challenge of evaluating local explanations for learning-to-rank models by proposing a systematic technique that uses tree-based models to extract ground truth feature importance scores via decision paths, and it finds that explanation accuracy varies significantly across models and data points on the MQ2008 dataset.
Local explanations of learning-to-rank (LTR) models are thought to extract the most important features that contribute to the ranking predicted by the LTR model for a single data point. Evaluating the accuracy of such explanations is challenging since the ground truth feature importance scores are not available for most modern LTR models. In this work, we propose a systematic evaluation technique for explanations of LTR models. Instead of using black-box models, such as neural networks, we propose to focus on tree-based LTR models, from which we can extract the ground truth feature importance scores using decision paths. Once extracted, we can directly compare the ground truth feature importance scores to the feature importance scores generated with explanation techniques. We compare two recently proposed explanation techniques for LTR models when using decision trees and gradient boosting models on the MQ2008 dataset. We show that the explanation accuracy in these techniques can largely vary depending on the explained model and even which data point is explained.