FedDRAW: Federated Dual Reputation Annealing Weighting for Heterogeneous Multi-Institutional Chest Radiograph Classification

Maryam Moradpour; Anne-Christin Hauschild

cs.LGSep 4, 2026

FedDRAW: Federated Dual Reputation Annealing Weighting for Heterogeneous Multi-Institutional Chest Radiograph Classification

Authors: Maryam Moradpour, Anne-Christin Hauschild

Organizations: Institute for Predictive Deep Learning in Medicine and Healthcare, Justus-Liebig University, Gießen, Germany · Institute for Medical Informatics, University Medical Center Göttingen, Germany

Abstract

Artificial intelligence models are promising for medical diagnosis, but they require large numbers of unbiased data, which in medicine are distributed across hospitals and cannot be centralized to protect patient privacy. Federated Learning (FL) addresses this, since hospitals train one shared diagnostic model while patient data remain local. Training proceeds in communication rounds, in which each hospital trains the shared model locally and returns it to the server for merging by weighted average. This aggregation weight determines whose institutional knowledge shapes the result. Federated averaging (FedAvg) sets it in proportion to local sample count, so a small but informative hospital is permanently assigned a small influence, andl argest clients could dominate the global model even when they are less informative. We propose Federated Dual Reputation Annealing Weighting (FedDRAW), a server-side aggregation method that combines a data-size prior with the cosine similarity between client and global parameters under two coupled annealing schedules. An inner schedule shifts client reputation from the size prior towards similarity. An outer, deferred annealing schedule on the softmax inverse temperature keeps the weighting selective in the early and middle rounds and relaxes it to uniformity at convergence. We evaluate FedDRAW on 12 simulated client-partition scenarios of two chest radiograph datasets (CheXpert and ChestMNIST), against seven federated baselines under identical local training settings. FedDRAW achieved the highest average rank among all eight methods under both AUC and the geometric mean (GM) of sensitivity and specificity, which a Friedman test with Nemenyi post-hoc analysis confirmed to be a statistically significant difference between the methods. Scheduling two signals, rather than fixing the weights by sample count alone, could enable less biased diagnostic models.

Figures & tables

	CHX-1		CHX-2		CHX-3		CHX-4
Algorithm	AUC	GM	AUC	GM	AUC	GM	AUC	GM
FedDRAW	86.67 $\pm$ 0.25	78.76 $\pm$ 0.18	87.46 $\pm$ 0.30	77.83 $\pm$ 0.81	86.64 $\pm$ 0.18	78.03 $\pm$ 0.91	85.57 $\pm$ 0.60	77.35 $\pm$ 0.13
FedAvg	84.98 $\pm$ 0.68	76.01 $\pm$ 1.31	85.77 $\pm$ 0.33	75.41 $\pm$ 1.32	85.71 $\pm$ 0.43	76.71 $\pm$ 1.80	85.76 $\pm$ 0.68	75.61 $\pm$ 0.27
FedNova	85.33 $\pm$ 0.75	76.68 $\pm$ 0.35	86.16 $\pm$ 0.24	76.40 $\pm$ 1.00	85.63 $\pm$ 0.78	77.03 $\pm$ 0.39	82.88 $\pm$ 3.15	76.20 $\pm$ 0.27
FedProx	85.40 $\pm$ 0.33	76.26 $\pm$ 1.46	86.00 $\pm$ 0.98	76.52 $\pm$ 0.41	85.89 $\pm$ 0.61	77.14 $\pm$ 1.01	85.10 $\pm$ 0.32	75.53 $\pm$ 0.38
MOON	83.86 $\pm$ 0.83	76.36 $\pm$ 0.64	84.93 $\pm$ 0.50	76.61 $\pm$ 0.42	83.66 $\pm$ 0.19	75.17 $\pm$ 0.66	82.95 $\pm$ 0.33	72.58 $\pm$ 1.50

Table 1

	CHM-1		CHM-2		CHM-3		CHM-4		CHM-5
Algorithm	AUC	GM	AUC	GM	AUC	GM	AUC	GM	AUC	GM
FedDRAW	81.78 $\pm$ 0.53	74.84 $\pm$ 0.29	82.91 $\pm$ 0.13	75.86 $\pm$ 0.40	84.18 $\pm$ 0.14	76.61 $\pm$ 0.34	82.67 $\pm$ 0.77	75.48 $\pm$ 0.42	83.86 $\pm$ 0.69	76.53 $\pm$ 0.37
FedAvg	80.13 $\pm$ 0.17	73.06 $\pm$ 0.16	80.14 $\pm$ 0.60	73.22 $\pm$ 0.45	84.10 $\pm$ 0.27	76.01 $\pm$ 0.28	81.69 $\pm$ 0.21	73.67 $\pm$ 0.41	83.97 $\pm$ 0.10	76.30 $\pm$ 0.14
FedNova	78.36 $\pm$ 1.87	71.77 $\pm$ 1.49	80.69 $\pm$ 0.12	73.26 $\pm$ 0.11	84.25 $\pm$ 0.21	75.62 $\pm$ 0.06	81.37 $\pm$ 0.34	74.18 $\pm$ 0.09	83.77 $\pm$ 0.13	76.24 $\pm$ 0.15
FedProx	80.33 $\pm$ 0.67	73.16 $\pm$ 0.28	80.96 $\pm$ 0.55	73.58 $\pm$ 0.52	84.13 $\pm$ 0.25	75.72 $\pm$ 0.25	81.22 $\pm$ 0.68	72.12 $\pm$ 1.04	84.22 $\pm$ 0.04	76.61 $\pm$ 0.06
MOON	80.37 $\pm$ 0.41	73.04 $\pm$ 0.50	78.60 $\pm$ 0.51	71.40 $\pm$ 0.34	80.03 $\pm$ 0.60	72.95 $\pm$ 0.59	79.51 $\pm$ 0.43	72.50 $\pm$ 0.37	82.27 $\pm$ 0.16	74.99 $\pm$ 0.23

Table 2

Scenario	FedDRAW	FedAvg	FedNova	FedProx	MOON	FedDyn	SCAFFOLD	FedAdp
CHX-1	1 / 1	6 / 6	4 / 3	3 / 5	7 / 4	2 / 7	8 / 8	5 / 2
CHX-2	1 / 1	5 / 6	2 / 4	3 / 3	6 / 2	7 / 7	8 / 8	4 / 5
CHX-3	1 / 1	3 / 4	5 / 3	2 / 2	7 / 6	4 / 7	8 / 8	6 / 5
CHX-4	2 / 1	1 / 4	6 / 3	3 / 5	5 / 6	7 / 8	8 / 7	4 / 2
CHX-5	1 / 1	4 / 3	5 / 6	3 / 5	2 / 2	7 / 7	8 / 8	6 / 4
CHX-6	1 / 1	4 / 3	2 / 4	3 / 5	6 / 7	7 / 6	8 / 8	5 / 2

Table 3

Explore similar work

Sep 9, 2026cs.LG

OmniMed-FL: A Robust Multimodal Federated Learning Framework for Clinical Diagnosis

Simultaneous assessment of medical imaging and patient records is often required in clinical diagnosis. However, standard machine learning algorithms cannot analyze these data types together. Meanwhile, compliance with HIPAA and GDPR can constrain centralized aggregation of sensitive patient data. This leaves a crucial void of secure fusion of visual and textual context across distant networks. Thus, we present OmniMed-FL, a controlled systems study of multimodal federated learning for five-class clinical condition classification (Normal, Pneumonia, COVID-19, Pleural Effusion, Cardiomegaly). Our proxy corpus pairs 3,000 public chest radiographs with 3,000 class-conditioned synthetic notes, matched by class, not by patient. The framework benchmarks eight fusion strategies, three initializations, four missing-text imputation rules, and matched federated baselines under non-IID Dirichlet partitioning across 3 to 20 hospital clients. As all notes are synthetic and pairing is not patient-level, these are descriptive proxy comparisons, not estimates of diagnostic performance or deployment readiness. Within those limits with clients (

K=5

) and severe skew (

α=0.1

), local-only training achieves a macro-F1 score of 0.297, FedAvg achieves

0.662\pm0.074

, FedProx

0.737\pm0.085

, a matched FedMME-style one-shot ensemble

0.647\pm0.080

, and our SCAFFOLD-AdamW adaptation

0.070\pm0.015

, the 0.075 FedProx-FedAvg gap falling inside the wider of the two two-seed standard deviations. Over a

4\times3

grid, label skew costs up to 0.27 F1 whereas a near-sevenfold client increase costs at most 0.10, while bidirectional volume grows linearly to 183.5 GiB at

K=20

. Multimodal fusion leads on both corpora, scoring 0.956 against 0.934 for text and 0.664 for images on the synthetic corpus and 0.906 against 0.880 and 0.737 on the radiograph corpus, for

2.3\times

the model state of text alone.

Ayush Debnath, Ruelia Saha, Sudip Misra

Indian Institute of Technology Kharagpur, India · KTH Royal Institute of Technology, Sweden · SRM University-AP, India +1

Multimodal Federated Learning Medical Diagnosis Non-IID Federated Learning Multimodal Classification Multimodal Fusion

Ask PDF

Jul 22, 2026cs.LG

One Round Is All You Need: Analytic Federated Learning for Task-Heterogeneous Multi-Label Medical Image Classification

Federated learning (FL) enables multiple clinical institutions to collaboratively train a shared disease classifier without centralizing patient data. In practice, however, each institution annotates only the pathologies within its area of expertise, so the federation operates under task heterogeneity: each client holds labels for a strict subset of the target disease categories while the remaining classes are entirely unobserved at that site. Existing gradient-based FL methods fail under this setting because they require hundreds of communication rounds to converge and because missing class labels introduce systematic false-negative bias that the model cannot correct without a principled mechanism. We propose an analytic federated learning framework for multi-label medical image classification under task heterogeneity. The proposed method replaces iterative gradient optimization with three closed-form operations: a balanced label projection that neutralizes class-imbalance bias by normalizing positive and negative contributions to equal total mass; a per-class absolute aggregation law that independently assembles the optimal ridge-regression classifier for each disease category from the sufficient statistics uploaded by its annotating clients; and an optional analytic pseudo-label refinement round that propagates missing-class knowledge from a confidence-filtered teacher classifier to non-annotating clients. The entire procedure requires at most two communication rounds, irrespective of the degree of task heterogeneity or the number of participating clients. Experiments on ChestXray14 under four progressively severe missing-class configurations demonstrate that the proposed method consistently outperforms the state-of-the-art federated multi-label method FedMLP by up to 18.44 BACC points and 13.24 AUC points, while reducing the communication.

Afsaneh Mahanipour, Hana Khamfroush

Department of Computer Science University of Kentucky Lexington, KY, USA

Chest X-Ray Classification Federated Learning Aggregation Medical Image Classification Federated Learning Heterogeneous Federated Learning

Ask PDF

May 28, 2025cs.LG

Inclusive Federated Learning Through Compliance-Weighted Noise Allocation in Healthcare AI

Background: Federated learning (FL) enables collaborative training of clinical AI models without centralizing patient data, but adoption is limited by privacy concerns, heterogeneous institutional compliance, and resource disparities; standard differential privacy (DP) applies uniform noise to all clients, penalizing well-compliant or under-resourced institutions. Objective: We introduce a compliance-aware FL framework that adapts DP to institutional compliance, letting lower-compliance sites participate without uniformly penalizing others. Methods: A compliance scoring tool aligned with HIPAA, GDPR, NIST, ISO, and HL7/FHIR maps each client score to a per-step Gaussian noise scale for server-side DP-SGD on a small aggregator dataset. The formal

(ε,δ)

bound applies to the aggregator dataset under a semi-honest aggregator; client-level DP needs secure aggregation (future work). We evaluate five FL strategies on PneumoniaMNIST and BreastMNIST (16 clients, 50 rounds, five seeds); the cumulative aggregator-dataset

ε

is about 1434 (Breast) and 513 (Pneumonia) at

δ=10^{-5}

. Results: Including 12 lower-compliance clients (Experiment 1) versus a compliant-only baseline (Experiment 4) changed BreastMNIST accuracy by +4.5 (FedAvg), +6.8 (FedMedian), +5.2 (FedProx), +1.6 (FedYogi), and -4.1 (FedAdam) percentage points (pooled +2.8 pp; not significant at n=5; up to +17 pp per configuration); compliance-weighted allocation matched uniform server-side DP at equal mean noise (+0.1 pp), carrying no utility penalty, and first-round noise cost 1.3 pp (Breast) and 2.5 pp (Pneumonia, FedAvg). Conclusions: Compliance-weighted server-side DP lets lower-compliance institutions join FL without degrading performance, giving auditable per-site noise control at no utility cost; formal guarantees apply to the aggregator dataset, with client-level DP requiring secure aggregation.

Santhosh Parampottupadam, Melih Coşğun, Sarthak Pati +7

German Cancer Research Center (DKFZ), Division of Medical Image Computing, Heidelberg, Germany · Department of Computer Engineering, Bilkent University, Universiteler 06800 Çankaya/Ankara, Türkiye · Medical Research Group, MLCommons, San Francisco, CA, USA +2

Federated Learning Aggregation Differentially Private Stochastic Gradient Descent Healthcare Privacy-Preserving ML Differentially Private Federated Learning Federated Learning+1

Ask PDF