Model-Free Counterfactual Subset Selection at Scale

TL;DR

This paper presents a scalable, model-free algorithm for selecting diverse, relevant counterfactual explanations directly from streaming data, enabling real-time, interpretable AI decisions without relying on synthetic data or full dataset storage.

Contribution

It introduces a novel streaming algorithm for counterfactual selection that is efficient, model-free, and maintains high-quality explanations in real-time environments.

Findings

Outperforms baseline methods on real-world and synthetic datasets.

Maintains $O("log k)$ update complexity per item.

Demonstrates robustness under adversarial conditions.

Abstract

Ensuring transparency in AI decision-making requires interpretable explanations, particularly at the instance level. Counterfactual explanations are a powerful tool for this purpose, but existing techniques frequently depend on synthetic examples, introducing biases from unrealistic assumptions, flawed models, or skewed data. Many methods also assume full dataset availability, an impractical constraint in real-time environments where data flows continuously. In contrast, streaming explanations offer adaptive, real-time insights without requiring persistent storage of the entire dataset. This work introduces a scalable, model-free approach to selecting diverse and relevant counterfactual examples directly from observed data. Our algorithm operates efficiently in streaming settings, maintaining $O(\log k)$ update complexity per item while ensuring high-quality counterfactual selection. Empirical evaluations on both real-world and synthetic datasets demonstrate superior performance over baseline methods, with robust behavior even under adversarial conditions.