# Heterogeneous Federated Learning with Prototype Alignment and Upscaling

arXiv: https://arxiv.org/abs/2507.04310 · HTML: https://mlsanity.com/paper/2507.04310/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 31 of its 45 references that are in the map.
- First posted on arXiv 6 Jul 2025.

## Builds on (most cited first, 30 of 31)

- [Communication-Efficient Learning of Deep Networks from Decentralized Data](https://mlsanity.com/paper/1602.05629/) (2016)
- [EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks](https://mlsanity.com/paper/1905.11946/) (2019)
- [ShuffleNet V2: Practical Guidelines for Efficient CNN Architecture Design](https://mlsanity.com/paper/1807.11164/) (2018)
- [CosFace: Large Margin Cosine Loss for Deep Face Recognition](https://mlsanity.com/paper/1801.09414/) (2018)
- [Ensemble Distillation for Robust Model Fusion in Federated Learning](https://mlsanity.com/paper/2006.07242/) (2020)
- [Deep Mutual Learning](https://mlsanity.com/paper/1706.00384/) (2017)
- [Personalized Federated Learning with Moreau Envelopes](https://mlsanity.com/paper/2006.08848/) (2020)
- [Federated Learning on Non-IID Data Silos: An Experimental Study](https://mlsanity.com/paper/2102.02079/) (2021)
- [FedMD: Heterogenous Federated Learning via Model Distillation](https://mlsanity.com/paper/1910.03581/) (2019)
- [Data-Free Knowledge Distillation for Heterogeneous Federated Learning](https://mlsanity.com/paper/2105.10056/) (2021)
- [FedProto: Federated Prototype Learning across Heterogeneous Clients](https://mlsanity.com/paper/2105.00243/) (2021)
- [HeteroFL: Computation and Communication Efficient Federated Learning for Heterogeneous Clients](https://mlsanity.com/paper/2010.01264/) (2020)
- [Think Locally, Act Globally: Federated Learning with Local and Global Representations](https://mlsanity.com/paper/2001.01523/) (2020)
- [Large-Margin Softmax Loss for Convolutional Neural Networks](https://mlsanity.com/paper/1612.02295/) (2016)
- [Personalized Cross-Silo Federated Learning on Non-IID Data](https://mlsanity.com/paper/2007.03797/) (2020)
- [Communication-Efficient On-Device Machine Learning: Federated Distillation and Augmentation under Non-IID Private Data](https://mlsanity.com/paper/1811.11479/) (2018)
- [FjORD: Fair and Accurate Federated Learning under heterogeneous targets with Ordered Dropout](https://mlsanity.com/paper/2102.13451/) (2021)
- [Hyperspherical Variational Auto-Encoders](https://mlsanity.com/paper/1804.00891/) (2018)
- [FedALA: Adaptive Local Aggregation for Personalized Federated Learning](https://mlsanity.com/paper/2212.01197/) (2022)
- [FedRolex: Model-Heterogeneous Federated Learning with Rolling Sub-Model Extraction](https://mlsanity.com/paper/2212.01548/) (2022)
- [Learning towards Minimum Hyperspherical Energy](https://mlsanity.com/paper/1805.09298/) (2018)
- [Tackling Data Heterogeneity in Federated Learning with Class Prototypes](https://mlsanity.com/paper/2212.02758/) (2022)
- [FedGH: Heterogeneous Federated Learning with Generalized Global Header](https://mlsanity.com/paper/2303.13137/) (2023)
- [Deep Hyperspherical Learning](https://mlsanity.com/paper/1711.03189/) (2017)
- [FedTGP: Trainable Global Prototypes with Adaptive-Margin-Enhanced Contrastive Learning for Data and Model Heterogeneity in Federated Learning](https://mlsanity.com/paper/2401.03230/) (2024)
- [Orthogonal Over-Parameterized Training](https://mlsanity.com/paper/2004.04690/) (2020)
- [Learning with Hyperspherical Uniformity](https://mlsanity.com/paper/2103.01649/) (2021)
- [Angular Visual Hardness](https://mlsanity.com/paper/1912.02279/) (2019)
- [Regularizing Neural Networks via Minimizing Hyperspherical Energy](https://mlsanity.com/paper/1906.04892/) (2019)
- [Hyperspherical Consistency Regularization](https://mlsanity.com/paper/2206.00845/) (2022)

## Abstract (arXiv)

Heterogeneity in data distributions and model architectures remains a significant challenge in federated learning (FL). Various heterogeneous FL (HtFL) approaches have recently been proposed to address this challenge. Among them, prototype-based FL (PBFL) has emerged as a practical framework that only shares per-class mean activations from the penultimate layer. However, PBFL approaches often suffer from suboptimal prototype separation, limiting their discriminative power. We propose Prototype Normalization (ProtoNorm), a novel PBFL framework that addresses this limitation through two key components: Prototype Alignment (PA) and Prototype Upscaling (PU). The PA method draws inspiration from the Thomson problem in classical physics, optimizing global prototype configurations on a unit sphere to maximize angular separation; subsequently, the PU method increases prototype magnitudes to enhance separation in Euclidean space. Extensive evaluations on benchmark datasets show that our approach better separates prototypes and thus consistently outperforms existing HtFL approaches. Notably, since ProtoNorm inherits the communication efficiency of PBFL and the PA is performed server-side, it is particularly suitable for resource-constrained environments.