Search papers, labs, and topics across Lattice.
This paper introduces a novel method for generating counterfactual explanations for time series classification by directly optimizing in the input space. The key innovation is the integration of soft-DTW alignment with k-NN from the target class to encourage realistic temporal structure in the generated counterfactuals. Experiments demonstrate that the proposed method achieves competitive validity and significantly outperforms existing approaches in distributional alignment with the target class, indicating superior temporal realism.
Time series counterfactual explanations can now be more realistic thanks to a novel soft-DTW-based approach that ensures temporal structure.
We present a new method for generating plausible counterfactual explanations for time series classification problems. The approach performs gradient-based optimization directly in the input space. To enforce plausibility, we integrate soft-DTW (dynamic time warping) alignment with $k$-nearest neighbors from the target class, which effectively encourages the generated counterfactuals to adopt a realistic temporal structure. The overall optimization objective is a multi-faceted loss function that balances key counterfactual properties. It incorporates losses for validity, sparsity, and proximity, alongside the novel soft-DTW-based plausibility component. We conduct an evaluation of our method against several strong reference approaches, measuring the key properties of the generated counterfactuals across multiple dimensions. The results demonstrate that our method achieves competitive performance in validity while significantly outperforming existing approaches in distributional alignment with the target class, indicating superior temporal realism. Furthermore, a qualitative analysis highlights the critical limitations of existing methods in preserving realistic temporal structure. This work shows that the proposed method consistently generates counterfactual explanations for time series classifiers that are not only valid but also highly plausible and consistent with temporal patterns.
