Federated-Continual Dynamic Segmentation of Histopathology guided by Barlow Continuity

2025-01-08

AI Summary

To address client shift in federated learning and catastrophic forgetting in continual learning for pathological image analysis, this paper proposes the Dynamic Barlow Continuity mechanism—the first framework unifying modeling of inter-institutional (spatial) and inter-temporal (temporal) distribution shifts. Under privacy-preserving constraints, the method guides client updates via a shared reference set, integrating federated learning, continual learning, a Barlow Twins–based self-supervised contrastive constraint, and dynamic consistency regularization, with Dice loss optimizing segmentation performance. On the BCSS and Semicol datasets, the proposed method reduces the Dice degradation caused by client shift from 15.8% to only 71.6% (i.e., improves absolute Dice by 55.8 points), and mitigates catastrophic forgetting—raising the retained performance from 42.5% to 62.8% (i.e., improves absolute Dice by 20.3 points). These results demonstrate substantial gains in both spatial and temporal generalization as well as model robustness.

Technology Category

• Machine Learning: Life-Long and Continual Learning
• Computer Vision: Segmentation
• Natural Language Processing: Safety and Robustness

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Security and Privacy: Data transparency and provenance

Abstract

Federated- and Continual Learning have been established as approaches to enable privacy-aware learning on continuously changing data, as required for deploying AI systems in histopathology images. However, data shifts can occur in a dynamic world, spatially between institutions and temporally, due to changing data over time. This leads to two issues: Client Drift, where the central model degrades from aggregating data from clients trained on shifted data, and Catastrophic Forgetting, from temporal shifts such as changes in patient populations. Both tend to degrade the model's performance of previously seen data or spatially distributed training. Despite both problems arising from the same underlying problem of data shifts, existing research addresses them only individually. In this work, we introduce a method that can jointly alleviate Client Drift and Catastrophic Forgetting by using our proposed Dynamic Barlow Continuity that evaluates client updates on a public reference dataset and uses this to guide the training process to a spatially and temporally shift-invariant model. We evaluate our approach on the histopathology datasets BCSS and Semicol and prove our method to be highly effective by jointly improving the dice score as much as from 15.8% to 71.6% in Client Drift and from 42.5% to 62.8% in Catastrophic Forgetting. This enables Dynamic Learning by establishing spatio-temporal shift-invariance.

Problem

• Federated Learning
• Continual Learning
• Catastrophic Forgetting

Innovation

• Dynamic Batch Continuity
• Federated Learning
• Catastrophic Forgetting