Hide&Seek: Learning to Explain in an End-to-End Differentiable Network
Abstract
Instance-wise feature selection is a valuable tool for interpreting labeled data and the predictions of black-box models. In contrast to global feature selection techniques, instance-wise methods dynamically identify important features for each instance. A growing number of methods learn a selector, which identifies important features, and a predictor, which uses these to make predictions. However, these pioneering methods face challenges including information leakage and lack of differentiability, which can slow training. In this paper, we present Hide&Seek, an end-to-end differentiable model for instance-wise feature selection. We jointly learn feature selection and prediction under a single objective without information leakage. Hide&Seek outperforms existing state-of-the-art models across a range of experiments and is fast to train. We achieve this by reformulating feature removal as a differentiable operation where instead of discretely removing features, we replace a proportion of each feature. Training is further stabilized via a parsimony-weight annealing framework.
Related
- FTU-Seek: Foundation Model-Guided Hard-Negative Learning for Sparse Functional Tissue Unit Segmentation
- Continuous and Differentiable Neural Network Representations of the Surfzone Sea-Surface from Lidar Point Clouds
- Differentiable Optimization Layers for Guaranteed Fairness in Deep Learning
- Difference-of-Convex Regularization for Graph Learning by Differentiable Programming
- Differentiable Hybrid Modelling for Learning and Optimising Chemical Transport Processes from Experimental Data
BibTeX
@misc{ellinson2026hideseek,
title = {{Hide&Seek: Learning to Explain in an End-to-End Differentiable Network}},
author = {Tal Ellinson and Hadi Mohasel Afshar and Sally Cripps},
howpublished = {arXiv},
year = {2026},
month = aug,
doi = {10.48550/arxiv.2608.16689},
eprint = {2608.16689},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2608.16689},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Hide&Seek: Learning to Explain in an End-to-End Differentiable Network
AU - Tal Ellinson
AU - Hadi Mohasel Afshar
AU - Sally Cripps
PY - 2026
DA - 2026/08/17
PB - arXiv
DO - 10.48550/arxiv.2608.16689
UR - https://arxiv.org/abs/2608.16689
AB - Instance-wise feature selection is a valuable tool for interpreting labeled data and the predictions of black-box models. In contrast to global feature selection techniques, instance-wise methods dynamically identify important features for each instance. A growing number of methods learn a selector, which identifies important features, and a predictor, which uses these to make predictions. However, these pioneering methods face challenges including information leakage and lack of differentiability, which can slow training. In this paper, we present Hide&Seek, an end-to-end differentiable model for instance-wise feature selection. We jointly learn feature selection and prediction under a single objective without information leakage. Hide&Seek outperforms existing state-of-the-art models across a range of experiments and is fast to train. We achieve this by reformulating feature removal as a differentiable operation where instead of discretely removing features, we replace a proportion of each feature. Training is further stabilized via a parsimony-weight annealing framework.
ER -