Federated Deconfounding and Debiasing Learning for Out-of-Distribution Generalization

TL;DR

This paper introduces FedDDL, a federated learning method that uses causal inference and counterfactual data to reduce bias and improve out-of-distribution generalization, achieving significant accuracy gains.

Contribution

It proposes a novel federated deconfounding and debiasing framework with causal graph analysis, counterfactual data generation, and causal prototypical regularization for OOD generalization.

Findings

Achieves 4.5% higher Top-1 accuracy on benchmark datasets.

Effectively reduces background bias in computer vision tasks.

Outperforms 9 state-of-the-art methods in OOD generalization.

Abstract

Attribute bias in federated learning (FL) typically leads local models to optimize inconsistently due to the learning of non-causal associations, resulting degraded performance. Existing methods either use data augmentation for increasing sample diversity or knowledge distillation for learning invariant representations to address this problem. However, they lack a comprehensive analysis of the inference paths, and the interference from confounding factors limits their performance. To address these limitations, we propose the \underline{Fed}erated \underline{D}econfounding and \underline{D}ebiasing \underline{L}earning (FedDDL) method. It constructs a structured causal graph to analyze the model inference process, and performs backdoor adjustment to eliminate confounding paths. Specifically, we design an intra-client deconfounding learning module for computer vision tasks to decouple background and objects, generating counterfactual samples that establish a connection between the background and any label, which stops the model from using the background to infer the label. Moreover, we design an inter-client debiasing learning module to construct causal prototypes to reduce the proportion of the background in prototype components. Notably, it bridges the gap between heterogeneous representations via causal prototypical regularization. Extensive experiments on 2 benchmarking datasets demonstrate that \methodname{} significantly enhances the model capability to focus on main objects in unseen data, leading to 4.5\% higher Top-1 Accuracy on average over 9 state-of-the-art existing methods.