FreqRISE: Explaining time series using frequency masking

Thea Brüsch, Kristoffer Knutsen Wickstrøm, Mikkel N. Schmidt, Tommy Sonne Alstrøm, Robert Jenssen
Proceedings of the 6th Northern Lights Deep Learning Conference (NLDL), PMLR 265:16-31, 2025.

Abstract

Time series data is fundamentally important for many critical domains such as healthcare, finance, and climate, where explainable models are necessary for safe automated decision-making. To develop explainable artificial intelligence in these domains therefore implies explaining salient information in the time series. Current methods for obtaining saliency maps assumes localized information in the raw input space. In this paper, we argue that the salient information of a number of time series is more likely to be localized in the frequency domain. We propose FreqRISE, which uses masking based methods to produce explanations in the frequency and time-frequency domain, and outperforms strong baselines across a number of tasks.

Cite this Paper

BibTeX
@InProceedings{pmlr-v265-brusch25a, title = {Freq{RISE}: Explaining time series using frequency masking}, author = {Br{\"u}sch, Thea and Wickstr{\o}m, Kristoffer Knutsen and Schmidt, Mikkel N. and Alstr{\o}m, Tommy Sonne and Jenssen, Robert}, booktitle = {Proceedings of the 6th Northern Lights Deep Learning Conference (NLDL)}, pages = {16--31}, year = {2025}, editor = {Lutchyn, Tetiana and Ramírez Rivera, Adín and Ricaud, Benjamin}, volume = {265}, series = {Proceedings of Machine Learning Research}, month = {07--09 Jan}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v265/main/assets/brusch25a/brusch25a.pdf}, url = {https://proceedings.mlr.press/v265/brusch25a.html}, abstract = {Time series data is fundamentally important for many critical domains such as healthcare, finance, and climate, where explainable models are necessary for safe automated decision-making. To develop explainable artificial intelligence in these domains therefore implies explaining salient information in the time series. Current methods for obtaining saliency maps assumes localized information in the raw input space. In this paper, we argue that the salient information of a number of time series is more likely to be localized in the frequency domain. We propose FreqRISE, which uses masking based methods to produce explanations in the frequency and time-frequency domain, and outperforms strong baselines across a number of tasks.} }
Endnote
%0 Conference Paper %T FreqRISE: Explaining time series using frequency masking %A Thea Brüsch %A Kristoffer Knutsen Wickstrøm %A Mikkel N. Schmidt %A Tommy Sonne Alstrøm %A Robert Jenssen %B Proceedings of the 6th Northern Lights Deep Learning Conference (NLDL) %C Proceedings of Machine Learning Research %D 2025 %E Tetiana Lutchyn %E Adín Ramírez Rivera %E Benjamin Ricaud %F pmlr-v265-brusch25a %I PMLR %P 16--31 %U https://proceedings.mlr.press/v265/brusch25a.html %V 265 %X Time series data is fundamentally important for many critical domains such as healthcare, finance, and climate, where explainable models are necessary for safe automated decision-making. To develop explainable artificial intelligence in these domains therefore implies explaining salient information in the time series. Current methods for obtaining saliency maps assumes localized information in the raw input space. In this paper, we argue that the salient information of a number of time series is more likely to be localized in the frequency domain. We propose FreqRISE, which uses masking based methods to produce explanations in the frequency and time-frequency domain, and outperforms strong baselines across a number of tasks.
APA
Brüsch, T., Wickstrøm, K.K., Schmidt, M.N., Alstrøm, T.S. & Jenssen, R.. (2025). FreqRISE: Explaining time series using frequency masking. Proceedings of the 6th Northern Lights Deep Learning Conference (NLDL), in Proceedings of Machine Learning Research 265:16-31 Available from https://proceedings.mlr.press/v265/brusch25a.html.

Related Material