Federated Causal Representation Learning in State-Space Systems for Decentralized Counterfactual Reasoning

2026-02-22

AI Summary

This work addresses the challenge of performing cross-client counterfactual reasoning in industrial asset networks, where client data are high-dimensional, private, and local models are immutable. The paper proposes the first federated causal representation learning framework: each client maps its observations into a low-dimensional latent state that disentangles intrinsic dynamics from control effects; the server aggregates these latent states to estimate global state transitions and control structures, enabling decentralized counterfactual inference without sharing raw data or modifying local models. Theoretical analysis establishes convergence guarantees and privacy preservation, while experiments on both synthetic and real-world industrial datasets demonstrate the method’s scalability, counterfactual accuracy, and performance closely approaching that of an ideal centralized model.

Technology Category

• Machine Learning: Causal Learning
• Reasoning under Uncertainty: Causality
• Knowledge Representation and Reasoning: Action, Change, and Causality

Application Category

• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web

Abstract

Networks of interdependent industrial assets (clients) are tightly coupled through physical processes and control inputs, raising a key question: how would the output of one client change if another client were operated differently? This is difficult to answer because client-specific data are high-dimensional and private, making centralization of raw data infeasible. Each client also maintains proprietary local models that cannot be modified. We propose a federated framework for causal representation learning in state-space systems that captures interdependencies among clients under these constraints. Each client maps high-dimensional observations into low-dimensional latent states that disentangle intrinsic dynamics from control-driven influences. A central server estimates the global state-transition and control structure. This enables decentralized counterfactual reasoning where clients predict how outputs would change under alternative control inputs at others while only exchanging compact latent states. We prove convergence to a centralized oracle and provide privacy guarantees. Our experiments demonstrate scalability, and accurate cross-client counterfactual inference on synthetic and real-world industrial control system datasets.

Problem

• federated learning
• causal representation
• counterfactual reasoning
• state-space systems
• decentralized inference

Innovation

• Federated Learning
• Causal Representation Learning
• State-Space Models
• Counterfactual Reasoning
• Decentralized Inference