Healthcare Classification on PubMed synthetic data (non-IID)
75.6AccuracyFederated Controllable Text Generation
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Federated Controllable Text GenerationBias=D, C, M, N, epsilon=inf, refinement=true2026.03 | 75.6 | 75.26 | |
| Federated Controllable Text GenerationBias=M, N, epsilon=inf, refinement=true2026.03 | 73.68 | 72.56 | |
| Federated Controllable Text GenerationBias=C, M, N, epsilon=inf, refinement=true2026.03 | 73.36 | 73.07 | |
| Federated Controllable Text GenerationBias=D, C, M, N, epsilon=8, refinement=true2026.03 | 72.42 | 71.35 | |
| Federated Controllable Text GenerationBias=D, C, M, N, epsilon=inf, refinement=false2026.03 | 71.36 | 71.41 | |
| Federated Controllable Text GenerationBias=C, M, N, epsilon=inf, refinement=false2026.03 | 70.84 | 70.89 | |
| Federated Controllable Text GenerationBias=C, M, N, epsilon=8, refinement=true2026.03 | 70.38 | 69.36 | |
| Federated Controllable Text GenerationBias=D, C, M, N, epsilon=8, refinement=false2026.03 | 67.31 | 66.42 | |
| Federated Controllable Text GenerationBias=M, N, epsilon=inf, refinement=false2026.03 | 66.52 | 66.05 | |
| Federated Controllable Text GenerationBias=C, M, N, epsilon=8, refinement=false2026.03 | 66.27 | 65.23 | |
| Federated Controllable Text GenerationBias=M, N, epsilon=8, refinement=true2026.03 | 64.37 | 63.25 | |
| Federated Controllable Text GenerationBias=M, N, epsilon=8, refinement=false2026.03 | 60.29 | 59.21 |