Time-series attribution maps with regularized contrastive learning

Steffen Schneider, Rodrigo González Laiz, Anastasiia Filippova, Markus Frey, Mackenzie W Mathis

TL;DR: We propose a regularized contrastive learning objective to estimate attribution maps on time series with identifiability guarantees, with applications in neuroscience

Abstract: Gradient-based attribution methods aim to explain decisions of deep learning models, but so far lack identifiability guarantees. Here, we propose a method to generate attribution maps with identifiability guarantees by developing a regularized contrastive learning algorithm (RegCL) trained on time-series data. We show theoretically that RegCL has favorable properties for identifying the Jacobian matrix of the data generating process. Empirically, we demonstrate robust approximation of zero vs. non-zero entries in the ground-truth attribution map on synthetic datasets, and significant improvements across previous attribution methods based on feature ablation, Shapley values, and other gradient-based methods. Our work constitutes a first example of identifiable inference of time-series attribution maps, and opens avenues better understanding of time-series data, such as for neural dynamics and decision-processes within neural networks.

Submission Number: 851