Jul 2026· ACM Transactions on Knowledge Discovery from Data· Vol 20, pp. 1 - 27· 0 citations· 49 references
Computer Science
TL;DR
PDKD solves the problem of model drift caused by the inconsistent distribution of distillation datasets and the local data by generating personalized distillation datasets for each client while protecting client data privacy.
Abstract
Knowledge Distillation (KD) is considered as an efficient way to replace the parameter averaging in federated learning, aiming to handle the clients with heterogeneous model architectures. Relying on the prepared distillation datasets across clients and the server, KD may encounter impractical difficulties in real-world implementations. Existing works explore the data-free KD in federated learning, which generates the distillation datasets on-site. However, the distillation datasets with global data distribution generated by these state-of-the-art schemes cannot be adapted to local non-IID data. In this article, we propose a new Personalized Data-Free Knowledge Distillation, namely PDKD, for federated learning under heterogeneous models and data. PDKD solves the problem of model drift caused by the inconsistent distribution of distillation datasets and the local data by generating personalized distillation datasets for each client while protecting client data privacy. In addition, we design a distillation dataset update scheme that maximizes the difference between teacher and client outputs on distillation datasets to accomplish deeper knowledge transfer. Furthermore, in order to accomplish the co-evolution of the teacher model and the clients’ model, PDKD incorporates a mutual distillation scheme. Numerous experiments show that PDKD significantly outperforms several state-of-the-art algorithms, with an 18% improvement in prediction accuracy and has a much lower communication cost than the compared algorithms.
A domain-aware proxy selection framework to better adopt proxy data for OOD problems is proposed and the experimental results show that the proposed models effectively address the challenges of distribution shifts under OOD with and without proxy data.
Federated Learning (FL) has demonstrated a promising future in privacy-friendly collaboration but it faces the data heterogeneity problem. Knowledge Distillation (KD) can serve as an effective method to address this issue. However, challenges arise from the unreliability of existing distillation methods in multi-domain scenarios. Prevalent distillation solutions primarily aim to fit the distributions of the global model directly by minimizing forward Kullback-Leibler divergence (KLD). This results in significant bias when the outputs of the global model are multi-peaked, which indicates the unreliability of distillation pathway . Mean-while, cross-domain update conflicts can notably reduce the accuracy of the global model (teacher model) in certain domains, reflecting the unreliability of the teacher model in these domains. In this work, we propose DKDR ( D ynamic K nowledge D istillation for R eliability in Federated Learning), which dynamically assigns weights to forward and reverse KLD based on knowledge discrepancies. This enables clients to fit the outputs from the teacher precisely. Moreover, we use knowledge decoupling to identify domain experts, thus clients can acquire reliable domain knowledge from experts. Empirical results from single-domain and multi-domain image classification tasks demonstrate the effectiveness of the proposed method and the efficiency of its key modules. The code is available at https://github.com/YueyangYuan/DKDR .
Yue Yuan, Wenke Huang, Frank Wan et al.· Neural Information Processin...· 1 citation
Federated learning (FL) is a practical framework that can train models on distributed user data while guaranteeing data privacy; however, due to heterogeneity in which each user has a different data distribution, problems frequently arise where both global and personalization performance deteriorate simultaneously. This dissertation presents methodologies for building efficient personalized models by identifying which strategies are effective in the global training stage and by showing how to preserve global knowledge while securing user-specific performance during local adaptation. First, we show that as data heterogeneity increases, the collapse of feature vectors is a more fundamental bottleneck than classifier weights, and propose a method that directly mitigates the discrepancy in representation magnitude between local and global models. Second, we analyze that a training approach that strengthens local alignment can induce forgetting of global knowledge (e.g., categories not observed locally), and propose a method that achieves both local alignment and global knowledge preservation by combining feature distillation based on the global model's feature vectors. Third, in federated personalized reward model learning with preference heterogeneity, we empirically verify the conventional belief that"increasing the number of global models yields better initialization,"and we show that when sufficient local fine-tuning is allowed, a single global initialization can instead provide stronger personalization performance. This study redefines the role of global initialization under data and preference heterogeneity and provides practical training strategies that simultaneously satisfy global knowledge preservation and personalization.
Conventional federated learning (FL) relies on parameter averaging, which forces clients to be doubly homogeneous: it demands an identical architecture and degrades under non-IID data. Real-world deployments usually break both assumptions. We sidestep both by building a decentralized knowledge distillation framework in which each client evaluates its peers'model snapshots on its own local data and distills from the resulting soft predictions. Because knowledge is transferred through the shared class posterior, clients are free to run different architectures; and because every teacher is evaluated on the student's own device, raw data never leaves the client, with no central server or public dataset required. Within this setting, we identify and address an under-examined problem: how to combine the peer teacher predictions. Existing methods, like uniform averaging, ignore how knowledge reliability varies across teachers and classes. We propose Class-wise Reliability-Aware Distillation (CRAD), which, per class, first discards teachers that disagree with the peer consensus and then takes a weighted average of the rest, weighting each teacher by its per-class reliability (precision, or inverse variance). Since the variance of an accuracy from $n$ samples scales as $1/n$, support enters automatically: among the teachers that survive filtering, a teacher is trusted for a class to the degree that it is both accurate and well-evidenced for it. On three image-classification benchmarks (CIFAR-10, CIFAR-100, and PathMNIST colon pathology), across heterogeneous architectures under severe non-IID skew, CRAD consistently outperforms competing methods in global accuracy.
Baraa Bilbeisi, Mengchen Fan, Baocheng Geng et al.· 1 citation
Federated Learning (FL) allows decentralized clients to train models collaboratively while preserving data privacy. However, distribution mismatch across clients often leads to poor global generalization and degraded local client-level performance. In such scenarios, some of the clients with their local models trained solely on local data may perform better than the globally learnt model, thus nullifying the benefits of collaborative federated learning. To address this, we propose SAPE-FL (Similarity-Aware Personalized Federated Learning), a novel personalization framework that anchors each client's model to both the global model and a similarity-weighted peer averaged model. By incorporating dynamic, client-specific regularization based on both model similarity and output similarity, SAPE-FL adaptively balances global knowledge transfer and peer collaboration while filtering out dissimilar clients. This dual anchoring mitigates negative transfer and enhances robustness in heterogeneous settings. We theoretically analyze our algorithm establishing its convergence guarantees and empirically show that SAPE-FL outperforms state-of-the-art methods under high statistical heterogeneity and low client data regimes.
A. Kumar, Sunil Gupta, Ngyuen Dang et al.· 0 citations
Local Inference Guided Aggregation for Heterogeneous Training Environments to Yield Enhancement Through Agreement and Regularization (LIGHTYEAR), a federated learning framework that performs update selection in function space using an NTK-based agreement score to characterize predictive behavior and determine a personalized aggregation set for each client.
Mirko Konstantin, S. Zachow, Anirban Mukhopadhyay· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
