Improving Class-Level Fairness Under Non-IID Data Distributions in Federated Learning
Bibliographic record
Abstract
Federated Learning (FL) has been gaining traction as a powerful solution for collaborative model training across decentralized devices, especially in privacy-sensitive domains. However, a persistent challenge in FL is the presence of non-IID data, where each participating device holds data that differs significantly from others. This uneven distribution creates an unfair training process, where classes that are rare across all devices receive little attention, while frequent classes dominate the model’s learning process. Such imbalance leads to biased global models that perform well on common data but poorly on rare or minority classes. This unfairness is especially problematic in real-world scenarios, such as healthcare or finance, where minority groups or rare conditions require equal attention. In this work, we propose a fairness-aware training framework designed to ensure that all classes, regardless of how common or rare they are, receive appropriate attention during training. By introducing a class-aware adjustment mechanism into the FL process, we ensure that rare classes are not overshadowed by more frequent ones. Our approach is simple, effective, and compatible with existing FL systems, making it a practical solution for promoting fairness in decentralized learning environments. Our simulations and experiments demonstrate that our proposed technique improves class-level fairness in FL models, while maintaining strong overall accuracy.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.014 | 0.042 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.002 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.002 | 0.002 |
| Scholarly communication | 0.002 | 0.004 |
| Open science | 0.003 | 0.004 |
| Research integrity | 0.002 | 0.003 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".