This paper proposes federated clustering with adaptive personalization (FedCAP), a parameter-efficient personalized FL framework that separates cluster-level representation learning from client-level adaptation.
Abstract
Federated learning (FL) enables cellular traffic prediction without centralizing raw base-station data, but statistical heterogeneity makes a single global model unsuitable for many clients. This paper proposes federated clustering with adaptive personalization (FedCAP), a parameter-efficient personalized FL framework that separates cluster-level representation learning from client-level adaptation. Clients are grouped using training-only daily traffic profiles, after which an LSTM backbone is trained by FedAvg within each cluster. Each client then freezes the cluster backbone and optimizes a residual bottleneck adapter locally. The adapter contains 4241 trainable parameters, 6.19% of the 68,483-parameter three-feature backbone and prediction head, and personalization transmits no model updates. In a shared-seed-42 comparison across 11 methods and four public datasets, FedCAP ranks first or second in 12 of 16 dataset–metric combinations. Across five shared seeds, its mean MAE is 6.66%, 7.47%, 2.64%, and 3.86% below FedAvg on the Milan, Trentino, Bihar, and Taiwan datasets, respectively. Holm-adjusted paired t-tests identify 6 significant dataset–metric differences, whereas exact Wilcoxon tests are not significant because each comparison contains only five nonzero seed-matched pairs; the statistical evidence is therefore interpreted conservatively.
Federated adaptation of time-series foundation models (TSFMs) is attractive for building energy forecasting because meter data are private, distributed, and highly non-IID. However, a single parameter-sharing strategy is unlikely to serve all pretrained TSFMs or building clients: fully shared adapters can suppress building-specific temporal behavior, while fully local adaptation discards cross-building transfer. We propose a personalized federated sparse adaptation framework with a heterogeneous temporal mixture-of-experts (MoE) adapter placed after the pretrained TSFM representation. A sequence-level router maps each 168-hour context window to a top-$k$ subset of experts specialized for periodicity, long-range interactions, local variation, trend-residual structure, and multi-resolution behavior. We compare global FL, local training, and personalized FL variants with globally shared or client-private expert banks. Across 50 buildings and three TSFM backbones, personalization consistently outperforms Global FL-MoE and Local MoE, while the best sparse-adaptation strategy varies by backbone and metric. Routing behavior further reveals client-level expert specialization, expert concentration, and near-uniform routing across backbones, showing that federated TSFM adaptation should be both client-aware and backbone-aware.
Priyanka Nihalchandani, Naman Srivastava, Varun Ojha 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
FedTraffic, a hierarchical federated learning framework for traffic flow forecasting that integrates Edge–Fog–Cloud computing, hybrid deep learning, adaptive federated optimization, and Explainable Artificial Intelligence, is proposed.
This work proposes FedHAttn, a novel hierarchical attention–based aggregation mechanism that explicitly models inter-client model feature importance to optimize global model performance and establishes an effective aggregator that balances accuracy, robustness, and efficiency in federated PM2.5 prediction.
Sudhir Kumar, Vaneet Kour, Shivendu Mishra et al.· International Journal of Mac...· 0 citations
Experiments under representative Non-IID settings on benchmark datasets show that PFLS-One achieves improved accuracy and faster convergence compared with representative baseline methods, and the convergence analysis under a non-convex objective provides theoretical support for the proposed method.
In multi-domain networking, virtual network function (VNF) scaling using machine learning requires an accurate prediction model while addressing privacy constraints and non-identical and independently distributed (non-IID) data across domains. Current models have used conventional federated learning (FL) methods, such as federated averaging (FedAvg), yet they suffer from degraded performance due to heterogeneous traffic patterns in multi-domain networks. However, existing studies have not addressed the impact of non-IID characteristics on FL-based VNF scaling or developed an effective solution to mitigate it. This paper proposes a metadata-clustering-driven FL method that clusters domains with different traffic patterns and trains cluster-specific models. We extract statistical, spectral, and temporal features to represent traffic disturbance. We apply principal component analysis (PCA) followed by K-means clustering to group time series. We apply FedAvg within clusters to train cluster-specific prediction models. To evaluate the performance of the proposed method, we set up a testbench to synchronize three non-IID patterns. The numerical results demonstrate that the proposed clustered FL method consistently achieves a lower mean squared error (MSE) than the FedAvg baseline across all four evaluated non-IID settings. The proposed method yields an MSE of 0.7056 (a 23.1% reduction from FedAvg’s 0.9176) under label skew, 0.3615 (a 4.6% reduction from 0.3790) under label and feature skew, 0.6958 (a 28.1% reduction from 0.9682) under label and quantity skew, and 0.3748 (a 0.7% reduction from 0.3774) under the combined skew setting. These consistent reductions in MSE demonstrate that the proposed method effectively mitigates the performance degradation typically caused by non-IID effects.
Run-Yu Wang, Eiji Oki· International Conference on...· 0 citations