CEPAE: Conditional Entropy-Penalized Autoencoders for Time Series Counterfactuals

2026-02-17

AI Summary

This work addresses counterfactual reasoning in time series affected by market events by proposing a novel method, CEPAE, which integrates structural causal models with autoencoders. Built upon the abduction–action–prediction framework, CEPAE introduces a conditional entropy penalty in the latent space—the first application of such a mechanism to encourage disentangled representation learning in temporal data. By combining variational and adversarial autoencoder architectures, the method significantly enhances the accuracy of counterfactual predictions. Extensive experiments on synthetic, semi-synthetic, and real-world datasets demonstrate that CEPAE consistently outperforms existing approaches across multiple evaluation metrics.

Technology Category

Machine Learning: Causal LearningReasoning under Uncertainty: CausalityKnowledge Representation and Reasoning: Action, Change, and Causality

Application Category

Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for rankingUser Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systemsGraph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs

Abstract

The ability to accurately perform counterfactual inference on time series is crucial for decision-making in fields like finance, healthcare, and marketing, as it allows us to understand the impact of events or treatments on outcomes over time. In this paper, we introduce a new counterfactual inference approach tailored to time series data impacted by market events, which is motivated by an industrial application. Utilizing the abduction-action-prediction procedure and the Structural Causal Model framework, we first adapt methods based on variational autoencoders and adversarial autoencoders, both previously used in counterfactual literature although not in time series settings. Then, we present the Conditional Entropy-Penalized Autoencoder (CEPAE), a novel autoencoder-based approach for counterfactual inference, which employs an entropy penalization loss over the latent space to encourage disentangled data representations. We validate our approach both theoretically and experimentally on synthetic, semi-synthetic, and real-world datasets, showing that CEPAE generally outperforms the other approaches in the evaluated metrics.

Problem

• counterfactual inference
• time series
• market events
• causal reasoning
• decision-making

Innovation

• counterfactual inference
• time series
• conditional entropy penalty
• disentangled representation
• autoencoder