FlowCF: Sparse Counterfactual Explanations for Mixed-Type Tabular Data using Flow Matching

2026-10-06

AI Summary

This study addresses the challenge that existing counterfactual explanation methods struggle to balance sparsity and proximity on mixed-type tabular data, often inducing excessive perturbations to numerical features. To overcome this limitation, this work proposes FlowCF, a model-agnostic generative framework based on flow matching that formulates counterfactual generation as sparse transport from factual instances to the target class. The method introduces hybrid flow operators to handle heterogeneous features and designs a gating network that leverages geometric properties to optimize sparsity. Evaluated across six benchmark datasets, FlowCF achieves state-of-the-art performance in both sparsity and proximity. Specifically, it modifies only 29% of numerical features while reducing displacement by 70%, significantly outperforming existing baselines.

Abstract

In the field of Explainable AI (XAI), counterfactual (CF) explanations interpret a model's decision by suggesting the changes to the input that would lead to a more favourable outcome. To be useful in practice, such an explanation should change few features and change them as little as possible, properties known as sparsity and proximity. We observe that existing methods remain limited in this respect, especially for numerical features, whether they are model-agnostic and amortised, or gradient-based with full access to the model. In this paper, we propose FlowCF, a model-agnostic generative method that frames CF generation as sparse transport from the factual to the target class. We solve this transport with flow matching, which we extend to mixed feature types with a novel mixed flow operator, and exploit the resulting geometry to optimise for sparsity through a gating network that minimises the number of features the transport changes. Extensive experiments on six benchmark datasets demonstrate that FlowCF produces the best numerical sparsity and proximity, changing 29% of the numerical features where the best baseline changes 89%, at 70% smaller displacement, while remaining comparable on the other desiderata.

Problem

• Counterfactual Explanations
• Explainable AI
• Sparsity
• Proximity
• Mixed-Type Tabular Data

Innovation

• Counterfactual Explanations
• Flow Matching
• Sparse Transport
• Mixed-Type Tabular Data
• Gating Network