ChaosMining: A Benchmark to Evaluate Post-Hoc Local Attribution Methods in Low SNR Environments
This work addresses the challenge of feature attribution in noisy real-world machine learning applications, though it appears incremental as it builds on existing RFE methods.
The study tackled the problem of evaluating post-hoc local attribution methods in low signal-to-noise ratio (SNR) environments by developing synthetic datasets across symbolic, image, and audio data, and introduced a novel extension to the recursive feature elimination (RFE) algorithm that showed strengths in prediction and feature selection but had scalability limitations.
In this study, we examine the efficacy of post-hoc local attribution methods in identifying features with predictive power from irrelevant ones in domains characterized by a low signal-to-noise ratio (SNR), a common scenario in real-world machine learning applications. We developed synthetic datasets encompassing symbolic functional, image, and audio data, incorporating a benchmark on the {\it (Model \(\times\) Attribution\(\times\) Noise Condition)} triplet. By rigorously testing various classic models trained from scratch, we gained valuable insights into the performance of these attribution methods in multiple conditions. Based on these findings, we introduce a novel extension to the notable recursive feature elimination (RFE) algorithm, enhancing its applicability for neural networks. Our experiments highlight its strengths in prediction and feature selection, alongside limitations in scalability. Further details and additional minor findings are included in the appendix, with extensive discussions. The codes and resources are available at \href{https://github.com/geshijoker/ChaosMining/}{URL}.