WinTSR: A Windowed Temporal Saliency Rescaling Method for Interpreting Time Series Deep Learning Models
This addresses the problem of interpreting time series deep learning models for researchers and practitioners, offering a novel method that improves upon existing limitations, though it is incremental in advancing interpretation techniques.
The authors tackled the challenge of interpreting complex time series forecasting models by introducing WinTSR, a method that captures temporal dependencies and rescales feature importance, which significantly outperformed 10 other interpretation techniques across 5 state-of-the-art models and 3 real-world datasets.
Interpreting complex time series forecasting models is challenging due to the temporal dependencies between time steps and the dynamic relevance of input features over time. Existing interpretation methods are limited by focusing mostly on classification tasks, evaluating using custom baseline models instead of the latest time series models, using simple synthetic datasets, and requiring training another model. We introduce a novel interpretation method, \textit{Windowed Temporal Saliency Rescaling (WinTSR)} addressing these limitations. WinTSR explicitly captures temporal dependencies among the past time steps and efficiently scales the feature importance with this time importance. We benchmark WinTSR against 10 recent interpretation techniques with 5 state-of-the-art deep-learning models of different architectures, including a time series foundation model. We use 3 real-world datasets for both time-series classification and regression. Our comprehensive analysis shows that WinTSR significantly outperforms other local interpretation methods in overall performance. Finally, we provide a novel, open-source framework to interpret the latest time series transformers and foundation models.