Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper introduces ConceptCF, a novel method for generating counterfactuals based on human-interpretable concepts in time series data, which enhances the explainability of AI models in critical fields like healthcare. By modifying meaningful concepts derived from time series decomposition, such as scale and frequency bands, ConceptCF provides clear explanations of how changes in these concepts can alter model predictions. Evaluations show that ConceptCF outperforms five leading methods across various metrics, establishing its effectiveness in producing valid and interpretable counterfactuals.
Counterfactuals generated through human-interpretable concepts can transform how we explain AI predictions in high-stakes applications, making them more actionable and understandable.
This paper proposes ConceptCF, a method for counterfactual generation that operates on human-interpretable concepts. In high-stakes domains such as healthcare and predictive maintenance, artificial intelligence models can increase efficiency and safety. Explainability is key to ensure these models rely on causal relationships rather than spurious correlations. Counterfactual explanations identify minimal modifications that would change a model's predictions. Existing methods for time series operate on individual points or subsequences without ensuring interpretability of the mutations. ConceptCF instead modifies meaningful concepts. As a result we can provide explanations in terms of these concepts, for example ``the model's prediction would be `Sit'instead of `Walk'if you increase the scale of the movement''. In this paper, the concepts are constructed through time series decomposition, resulting in concepts such as scale, and frequency bands. Counterfactuals are generated using a genetic algorithm that optimizes the concept mutations. Evaluation against five state-of-the-art approaches demonstrates that ConceptCF consistently achieves top-tier performance across validity, confidence, proximity, sparsity and plausibility metrics.