Abstract
Saliency methods provide visual explainability for deep image processing models by highlighting informative regions in the input images based on feature-wise (pixels) importance scores. These methods have been adopted to the time series domain, aiming to highlight important temporal regions in a sequence. This paper identifies, for the first time, the systematic failure of such methods in the time series domain when underlying patterns (e.g., dominant frequency or trend) are based on latent information rather than temporal regions. The latent feature importance postulation is highly relevant for the medical domain as many medical signals, such as EEG signals or sensor data for gate analysis, are commonly assumed to be related to the frequency domain. To the best of our knowledge, no existing post-hoc explainability method can highlight influential latent information for a classification problem. Hence, in this paper, we frame and analyze the problem of latent feature saliency detection. We first assess the explainability quality of multiple state-of-the-art saliency methods (Integrated Gradients, DeepLift, Kernel SHAP, Lime) on top of various classification methods (LSTM, CNN, LSTM and CNN trained via saliency-guided training) using simulated time series data with underlying temporal or latent space patterns. In conclusion, we identify that Integrated Gradients and DeepLift, if redesigned, could be potential candidates for latent saliency scores.
M. Schröder and A. Zamanian—Authors contributed equally.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Akiba, T., Sano, S., Yanase, T., Ohta, T., Koyama, M.: Optuna: a next-generation hyperparameter optimization framework. In: Proceedings of the 25st ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD 2019), Anchorage, USA (2019)
Ates, E., Aksar, B., Leung, V.J., Coskun, A.K.: Counterfactual explanations for multivariate time series. In: Proceedings of the 2021 International Conference on Applied Artificial Intelligence (ICAPAI), Halden, Norway, pp. 1–8 (2021)
Bach, S., Binder, A., Montavon, G., Klauschen, F., Müller, K., Samek, W.: On pixel-wise explanations for non-linear classifier decisions by layer-wise relevance propagation. PLoS One 10(7) (2015). https://doi.org/10.1371/journal.pone.0130140
Bastings, J., Filippova, K.: The elephant in the interpretability room: why use attention as explanation when we have saliency methods? In: Proceedings of the 2020 EMNLP Workshop BlackboxNLP: Analyzing and Interpreting Neural Networks for NLP (2020)
Bracewell, R.N.: The Fourier Transform and Its Applications, 3rd edn. McGraw-Hill, New York (2000)
Carrillo, A., Cantú, L.F., Noriega, A.: Individual explanations in machine learning models: a survey for practitioners. arXiv (2021). https://doi.org/10.48550/arXiv.2104.04144
Charte, D., Charte, F., del Jesus, M.J., Herrera, F.: An analysis on the use of autoencoders for representation learning: fundamentals, learning task case studies, explainability and challenges. Neurocomputing 404, 93–107 (2020). https://doi.org/10.1016/j.neucom.2020.04.057
Datta, A., Sen, S., Zick, Y.: Algorithmic transparency via quantitative input influence: theory and experiments with learning systems. In: Proceedings of the 2016 IEEE Symposium on Security and Privacy (SP), San Jose, USA, pp. 598–617 (2016)
Delaney, E., Greene, D., Keane, M.T.: Instance-based counterfactual explanations for time series classification. In: Sánchez-Ruiz, A.A., Floyd, M.W. (eds.) ICCBR 2021. LNCS (LNAI), vol. 12877, pp. 32–47. Springer, Cham (2021). https://doi.org/10.1007/978-3-030-86957-1_3
Falcon, W.: Pytorch lightning (2019). https://doi.org/10.5281/zenodo.3828935
Fong, R.C., Vedaldi, A.: Interpretable explanations of black boxes by meaningful perturbation. In: Proceedings of the 2017 IEEE International Conference on Computer Vision (ICCV), Venice, Italy, pp. 3449–3457 (2017)
Geweke, J.F., Singleton, K.J.: Latent variable models for time series: a frequency domain approach with an application to the permanent income hypothesis. J. Econometrics 17, 287–304 (1981). https://doi.org/10.1016/0304-4076(81)90003-8
Guidotti, R., Monreale, A., Ruggieri, S., Turini, F., Giannotti, F., Pedreschi, D.: A survey of methods for explaining black box models. ACM Comput. Surv. 51(5) (2018). https://doi.org/10.1145/3236009
Guidotti, R., Monreale, A., Spinnato, F., Pedreschi, D., Giannotti, F.: Explaining any time series classifier. In: Proceedings of the 2020 IEEE Second International Conference on Cognitive Machine Intelligence (CogMI), Atlanta, USA, pp. 167–176 (2020)
Hochreiter, S., Schmidhuber, J.: Long short-term memory. Neural Comput. 9(8), 1735–1780 (1997)
Ismail, A.A., Corrada Bravo, H., Feizi, S.: Improving deep learning interpretability by saliency guided training. In: Proceedings of the 35th Conference on Neural Information Processing Systems (NeurIPS 2021), pp. 26726–26739 (2021)
Ismail, A.A., Gunady, M.K., Corrada Bravo, H., Feizi, S.: Benchmarking deep learning interpretability in time series predictions. In: Proceedings of the 34th Conference on Neural Information Processing Systems (NeurIPS), pp. 6441–6452 (2020)
Karlsson, I., Rebane, J., Papapetrou, P., Gionis, A.: Locally and globally explainable time series tweaking. Knowl. Inf. Syst. 62(5), 1671–1700 (2019). https://doi.org/10.1007/s10115-019-01389-4
Kokhlikyan, N., et al.: Captum: a unified and generic model interpretability library for PyTorch. arXiv (2020). https://doi.org/10.48550/arXiv.2009.07896
Le Cun, Y., et al.: Handwritten digit recognition: applications of neural network chips and automatic learning. IEEE Commun. Mag. 27(11), 41–46 (1989). https://doi.org/10.1109/35.41400
Lim, B., Arik, S., Loeff, N., Pfister, T.: Temporal fusion transformers for interpretable multi-horizon time series forecasting. Int. J. Forecast. 37, 1748–1764 (2021). https://doi.org/10.1016/j.ijforecast.2021.03.012
Lipovetsky, S., Conklin, M.: Analysis of regression in game theory approach. Appl. Stoch. Model. Bus. Ind. 17, 319–330 (2001). https://doi.org/10.1002/asmb.446
Loeffler, C., Lai, W.C., Eskofier, B., Zanca, D., Schmidt, L., Mutschler, C.: Don’t get me wrong: how to apply deep visual interpretations to time series. arXiv (2022). https://doi.org/10.48550/ARXIV.2203.07861
Lundberg, S.M., Lee, S.I.: A unified approach to interpreting model predictions. In: Proceedings of the 31st International Conference on Neural Information Processing Systems (NIPS 2017), Long Beach, USA, pp. 4768–4777 (2017)
Mikolov, T., Sutskever, I., Chen, K., Corrado, G.S., Dean, J.: Distributed representations of words and phrases and their compositionality. In: Proceedings of the 27th Conference on Neural Information Processing Systems (NeurIPS 2013), Lake Tahoe, USA (2013)
Montavon, G., Lapuschkin, S., Binder, A., Samek, W., Müller, K.R.: Explaining nonlinear classification decisions with deep Taylor decomposition. Pattern Recogn. 65, 211–222 (2017). https://doi.org/10.1016/j.patcog.2016.11.008
Neely, M., Schouten, S.F., Bleeker, M.J.R., Lucic, A.: Order in the court: explainable AI methods prone to disagreement. In: Proceedings of the ICML Workshop on Theoretic Foundation, Criticism, and Application Trend of Explainable AI (2021)
Parvatharaju, P.S., Doddaiah, R., Hartvigsen, T., Rundensteiner, E.A.: Learning saliency maps to explain deep time series classifiers. In: Proceedings of the 30th ACM International Conference on Information & Knowledge Management, Virtual Event, Queensland, Australia, pp. 1406–1415 (2021)
Paszke, A., et al.: PyTorch: an imperative style, high-performance deep learning library. In: Proceedings of the 32nd Conference on Neural Information Processing Systems (NeurIPS 2019), Vancouver, Canada, pp. 8024–8035 (2019)
Petsiuk, V., Das, A., Saenko, K.: Rise: randomized input sampling for explanation of black-box models. In: Proceedings of the 29th British Machine Vision Conference (BMVC), Newcastle, UK (2018)
Ribeiro, M.T., Singh, S., Guestrin, C.: “Why should I trust you?” explaining the predictions of any classifier. In: Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD 2016), New York, USA, pp. 1135–1144 (2016)
Schlegel, U., Keim, D.A.: Time series model attribution visualizations as explanations. In: Proceedings of the 2021 IEEE Workshop on TRust and EXpertise in Visual Analytics (TREX), New Orleans, USA, pp. 27–31 (2021)
Schlegel, U., Oelke, D., Keim, D.A., El-Assady, M.: An empirical study of explainable AI techniques on deep learning models for time series tasks. In: Proceedings of the Pre-registration Workshop NeurIPS 2020, Vancouver, Canada (2020)
Shapley, L.S.: A Value for N-Person Games, pp. 307–317. Princeton University Press, Princeton (1953)
Shrikumar, A., Greenside, P., Kundaje, A.: Learning important features through propagating activation differences. In: Proceedings of the 34th International Conference on Machine Learning (ICML 2017), Sydney, Australia, vol. 70 (2017)
Shrikumar, A., Greenside, P., Shcherbina, A., Kundaje, A.: Not just a black box: learning important features through propagating activation differences. In: Proceedings of the 33rd International Conference on Machine Learning (ICML 2016), New York, USA, vol. 48 (2016)
Simonyan, K., Vedaldi, A., Zisserman, A.: Deep inside convolutional networks: visualising image classification models and saliency maps. arXiv (2014). https://doi.org/10.48550/arXiv.1312.6034
Smilkov, D., Thorat, N., Kim, B., Kim, B., Viégas, F.B., Wattenberg, M.: SmoothGrad: removing noise by adding noise. arXiv (2017). https://doi.org/10.48550/arXiv.1706.03825
Springenberg, J.T., Dosovitskiy, A., Brox, T., Riedmiller, M.A.: Striving for simplicity: the all convolutional net. arXiv (2015). https://doi.org/10.48550/arXiv.1412.6806
Sundararajan, M., Taly, A., Yan, Q.: Axiomatic attribution for deep networks. In: Proceedings of the 34th International Conference on Machine Learning (ICML 2017), Sydney, Australia, pp. 3319–3328 (2017)
Štrumbelj, E., Kononenko, I.: Explaining prediction models and individual predictions with feature contributions. Knowl. Inf. Syst. 41(3), 647–665 (2014)
Wang, Z., Samsten, I., Mochaourab, R., Papapetrou, P.: Learning time series counterfactuals via latent space representations. In: Soares, C., Torgo, L. (eds.) DS 2021. LNCS (LNAI), vol. 12986, pp. 369–384. Springer, Cham (2021). https://doi.org/10.1007/978-3-030-88942-5_29
Wiegreffe, S., Pinter, Y.: Attention is not not explanation. In: Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP), Hong Kong, China, pp. 11–20 (2019)
Ye, L., Keogh, E.: Time series shapelets: a new primitive for data mining. In: Proceedings of the 15th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD 2009), Paris, France, pp. 947–956 (2009)
Zeiler, M.D., Fergus, R.: Visualizing and understanding convolutional networks. In: Fleet, D., Pajdla, T., Schiele, B., Tuytelaars, T. (eds.) ECCV 2014. LNCS, vol. 8689, pp. 818–833. Springer, Cham (2014). https://doi.org/10.1007/978-3-319-10590-1_53
Acknowledgments
We thank Oleksandr Zadorozhnyi for his valuable support throughout the course of the research project. We thank Ruijie Chen, Elisabeth Pachl and Adrian Schwaiger for proofreading the manuscript and providing instructive feedback.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
A Appendix
A Appendix
1.1 A.1 Synthetic Data Generation
Based on the Fourier series latent model, a time series \(x_t, t=1,...,T\) is modeled as
To simulated data, let \(\tilde{n}\) represent the number of amplitudes present in the series, i.e. \(\forall i > \tilde{n}, A_i = 0\). For simplicity, we consider centered stationary periodic time series in the data generation process, i.e. \(a_0 = 0\). In this case, the value at every time step t is calculated as
We refer to the notions amplitude A, frequency \(\omega \), phase shift \(\phi \) as concepts. The separate Fourier coefficients \(A_i, \omega _i, \phi _i\) for \(i=1,...,\tilde{T}\) are referred to as latent features. The latent features frequency \(\omega _i\) and phase shift \(\phi _i\) are each sampled from a uniform distribution. The sampling intervals are chosen with respect to the specific intention in the experiment design. To simulate the amplitude parameters \(A_i\), a dominant amplitude \(A_1\) is sampled. The next amplitudes are calculated considering an exponential decay with a fixed rate dec:
This makes the first frequency i.e. \(\omega _1\) to be the dominant frequency of the Fourier series. Throughout the experiments, all time series were generated with an equal length of 300 time steps. i.e. \(T=300\).
For assigning class labels to the time series samples, we consider the following two scenarios.
Scenario 1: Label based on the presence of a shapelet
For assigning shape-based labels to the time series, a shapelet is inserted at a random or fixed position into all time series \(X \in D\) belonging to one class. The shapelet is a second simulated Fourier series of length \(l \le T\), where \(l = \text {window-ratio} \cdot T\) for a chosen window ratio. We define the sampling intervals for the latent features of the shapelet to be non-intersecting with the sampling intervals of the latent features of the original time series X. The resulting shapelet replaces the original time series in the interval \([j, j+l]\), where
Scenario 2: Label based on differences in the latent features
Following the investigation of the effectiveness of explainability methods for latent features, we introduce a second simulation scenario where the labels depend on a difference in the sampling distribution of latent features of the time series. This scenario highlights the main focus of this project and represents our novel view of explainability methods for time series. Similar to the first scenario, the time series are sampled as discretized Fourier series with latent variables \(\omega , A\) and \(\phi \). The latent dependency is induced as follows:
-
1.
Two normal distributions with different means (based on Table 3) are selected for classes 0 and 1. For positive parameters, the distributions are log-normal.
-
2.
Per each class, N/2 Fourier parameters are sampled from the given distributions.
-
3.
The rest of the parameters are sampled from the same distribution for both classes.
-
4.
Sampled parameters are given to the deterministic Fourier series in Eq. 1 to generate the temporal samples. Rows are then labeled with the associated class, from the corresponding distribution of which the informative parameters are sampled.
1.2 A.2 Data Set Description
Based on the data generation method described above, we design ten different mechanisms for binary classification of univariate time series. Table 2 lists the parameters and algorithms for assigning labels to each sample. In Table 3 the parameters used for sampling the Fourier series are presented. The complete simulation code base can be found in the GitHub repository at https://github.com/m-schroder/TSExplainability.
Rights and permissions
Copyright information
© 2023 The Author(s), under exclusive license to Springer Nature Switzerland AG
About this paper
Cite this paper
Schröder, M., Zamanian, A., Ahmidi, N. (2023). Post-hoc Saliency Methods Fail to Capture Latent Feature Importance in Time Series Data. In: Chen, H., Luo, L. (eds) Trustworthy Machine Learning for Healthcare. TML4H 2023. Lecture Notes in Computer Science, vol 13932. Springer, Cham. https://doi.org/10.1007/978-3-031-39539-0_10
Download citation
DOI: https://doi.org/10.1007/978-3-031-39539-0_10
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-031-39538-3
Online ISBN: 978-3-031-39539-0
eBook Packages: Computer ScienceComputer Science (R0)