FeDMRA: Federated Incremental Learning with Dynamic Memory Replay Allocation
About
In federated healthcare systems, Federated Class-Incremental Learning (FCIL) has emerged as a key paradigm, enabling continuous adaptive model learning among distributed clients while safeguarding data privacy. However, in practical applications, data across agent nodes within the distributed framework often exhibits non-independent and identically distributed (non-IID) characteristics, rendering traditional continual learning methods inapplicable. To address these challenges, this paper covers more comprehensive incremental task scenarios and proposes a dynamic memory allocation strategy for exemplar storage based on the data replay mechanism. This strategy fully taps into the inherent potential of data heterogeneity, while taking into account the performance fairness of all participating clients, thereby establishing a balanced and adaptive solution to mitigate catastrophic forgetting. Unlike the fixed allocation of client exemplar memory, the proposed scheme emphasizes the rational allocation of limited storage resources among clients to improve model performance. Furthermore, extensive experiments are conducted on three medical image datasets, and the results demonstrate significant performance improvements compared to existing baseline models.
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Federated Class-and-Domain-Incremental Learning | FCDIL | Aavg85.26 | 14 | |
| Federated Class-Incremental Learning | FCIL-M | Avg Acc83.06 | 14 | |
| Federated Class-Incremental Learning | FCIL-A | Aavg83.26 | 14 | |
| Federated Class-Incremental Learning | FCIL-H | Average Accuracy (Aavg)91.77 | 14 | |
| Federated Domain-Incremental Learning | FDIL | Aavg82.71 | 14 | |
| Federated Class-Domain Incremental Learning | FCDIL alpha=0.5 (test) | Average Accuracy (Aavg)83.57 | 7 | |
| Federated Class-Incremental Learning | FCIL-A alpha=0.5 (test) | Average Accuracy (Aavg)80 | 7 | |
| Federated Class-Incremental Learning | FCIL-H alpha=0.5 (test) | Average Accuracy88.48 | 7 | |
| Federated Class-Incremental Learning | FCIL-M alpha=0.5 (test) | Average Accuracy77.29 | 7 | |
| Federated Domain-Incremental Learning | FDIL alpha=0.5 (test) | Average Accuracy (Aavg)75.83 | 7 |