DeepROCK: Error-controlled interaction detection in deep neural networks
Primary Area: visualization or interpretation of learned representations
Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.
Keywords: Interpretable AI, Explainable AI, interaction detection, FDR, reproducibility
Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2024/AuthorGuide.
TL;DR: The first paper to detect interactions from deep neural networks subject to error control
Abstract: The complexity of deep neural networks (DNNs) makes them powerful but also makes them challenging to interpret, hindering their applicability in error-intolerant domains. Existing methods attempt to reason about the internal mechanism of DNNs by identifying feature interactions that influence prediction outcomes. However, such methods typically lack a systematic strategy to prioritize interactions while controlling confidence levels, making them difficult to apply in practice for scientific discovery and hypothesis validation. In this paper, we introduce a method, called DeepROCK, to address this limitation by using knockoffs, which are dummy variables that are designed to mimic the dependence structure of a given set of features while being conditionally independent of the response. Together with a novel DNN architecture involving a pairwise-coupling layer, DeepROCK jointly controls the false discovery rate (FDR) and maximizes statistical power. In addition, we identify a challenge in correctly controlling FDR using off-the-shelf feature interaction importance measures. DeepROCK overcomes this challenge by proposing a calibration procedure applied to existing interaction importance measures to make the FDR under control at a target level. Finally, we validate the effectiveness of DeepROCK through extensive experiments on simulated and real datasets.
Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors' identity.
No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.
Submission Number: 7743
Loading