Interpretability in Deep Time Series Models Demands Semantic Alignment
Abstract
Deep time series models continue to improve predictive performance, yet their deployment remains limited by their black-box nature. In response, existing interpretability approaches in the field keep focusing on explaining the internal model computations, without addressing whether they align or not with how a human would reason about the studied phenomenon. Instead, we state interpretability in deep time series models should pursue semantic alignment: predictions should be expressed in terms of variables that are meaningful to the end user, mediated by spatial and temporal mechanisms that admit user-dependent constraints. In this paper, we formalize this requirement and require that, once established, semantic alignment must be preserved under temporal evolution: a constraint with no analog in static settings. Provided with this definition, we outline a blueprint for semantically aligned deep time series models, identify properties that support trust, and discuss implications for model design.
Keywords
Cite
@article{arxiv.2602.02239,
title = {Interpretability in Deep Time Series Models Demands Semantic Alignment},
author = {Giovanni De Felice and Riccardo D'Elia and Alberto Termine and Pietro Barbiero and Giuseppe Marra and Silvia Santini},
journal= {arXiv preprint arXiv:2602.02239},
year = {2026}
}