Journal article · 2026 · Open access

FedPKD + : A Prototype-Based Framework for Efficient and Flexible Heterogeneous Federated Learning

Tong Liu, Feng Lyu, Shucheng Li, Cheng Tang, Yongheng Deng, Conghao Zhou, Huaqing Wu

IEEE Transactions on Mobile Computing, 25(11), 20056–20073 · Published 7 July 2026 · DOI 10.1109/tmc.2026.3710770

Open full textCite this paper

In brief

“FedPKD + : A Prototype-Based Framework for Efficient and Flexible Heterogeneous Federated Learning” is a 2026 journal article by Tong Liu and 6 co-authors, published in IEEE Transactions on Mobile Computing. OpenAlex counts 0 citations as of 19 November 2026. It is open access under a CC BY licence: the full text is free to read at IEEE Transactions on Mobile Computing.

Citations
0
References
53
Authors
7

Abstract

Federated Learning (FL) enables collaborative model training across mobile and edge devices without sharing raw data, but its deployment is hindered bysystem heterogeneityandnon-IIDdata. Existing FL methods either require homogeneous architectures or suffer from accuracy loss and high communication overhead in heterogeneous settings. To address these issues, we propose FedPKD+, a prototype-based knowledge distillation framework that leverages both output-space knowledge (logits) and feature-space knowledge (prototypes) to enhance flexibility and efficiency. Preliminary experiments reveal three key directions for improvement: logit quality, public data utilization, and feature regularization. FedPKD$^+$realizes them with four interacting modules: (1)dual knowledge transfershares logits for supervision and prototypes for feature regularization, (2)heterogeneous prototype alignmentmakes cross-model prototypes aggregatable, (3)prototype-based ensemble distillation and data filteringuse quality-weighted logits and global prototypes to refine server knowledge and filter public samples, and (4)server knowledge transferfeeds the refined knowledge back to clients Beyond framework design, we provide a convergence analysis of prototype-based KD in FL, proving both single-round and cross-round convergence under time-varying objectives. Extensive experiments under diverse non-IID settings demonstrate that FedPKD+consistently outperforms state-of-the-art baselines in both learning performance and communication efficiency.

Cite this paper

Formatted from the catalogue record. Check names, capitals and page numbers against the paper before you submit.

APA 7

Liu, T., Lyu, F., Li, S., Tang, C., Deng, Y., Zhou, C., & Wu, H. (2026). FedPKD + : A Prototype-Based Framework for Efficient and Flexible Heterogeneous Federated Learning. IEEE Transactions on Mobile Computing, 25(11), 20056–20073. https://doi.org/10.1109/tmc.2026.3710770

MLA 9

Liu, Tong, et al. “FedPKD + : A Prototype-Based Framework for Efficient and Flexible Heterogeneous Federated Learning.” IEEE Transactions on Mobile Computing, vol. 25, no. 11, 2026, pp. 20056–20073. https://doi.org/10.1109/tmc.2026.3710770.

Chicago

Liu, Tong, Feng Lyu, Shucheng Li, Cheng Tang, Yongheng Deng, Conghao Zhou, and Huaqing Wu. 2026. “FedPKD + : A Prototype-Based Framework for Efficient and Flexible Heterogeneous Federated Learning.” IEEE Transactions on Mobile Computing 25 (11): 20056–20073. https://doi.org/10.1109/tmc.2026.3710770.

Harvard

Liu, T. et al. (2026) ‘FedPKD + : A Prototype-Based Framework for Efficient and Flexible Heterogeneous Federated Learning’, IEEE Transactions on Mobile Computing, 25(11), pp. 20056–20073. doi:10.1109/tmc.2026.3710770.

IEEE

T. Liu et al., “FedPKD + : A Prototype-Based Framework for Efficient and Flexible Heterogeneous Federated Learning,” IEEE Transactions on Mobile Computing, vol. 25, no. 11, pp. 20056–20073, 2026, doi: 10.1109/tmc.2026.3710770.

BibTeX

@article{liu2026fedpkd,
  title     = {{FedPKD + : A Prototype-Based Framework for Efficient and Flexible Heterogeneous Federated Learning}},
  author    = {Liu, Tong and Lyu, Feng and Li, Shucheng and Tang, Cheng and Deng, Yongheng and Zhou, Conghao and Wu, Huaqing},
  journal   = {IEEE Transactions on Mobile Computing},
  year      = {2026},
  volume    = {25},
  number    = {11},
  pages     = {20056--20073},
  doi       = {10.1109/tmc.2026.3710770},
  url       = {https://doi.org/10.1109/tmc.2026.3710770}
}

Download for a reference manager: BibTeX (.bib) · RIS for Zotero, Mendeley and EndNote (.ris)

Topic and keywords

Physical Sciences › Computer Science › Artificial Intelligence › Privacy-Preserving Technologies in Data

  • federated learning
  • knowledge distillation
  • non-IID data
  • system heterogeneity
  • prototype alignment
  • ensemble distillation
  • data filtering
  • convergence analysis

Work with this paper

Most-cited papers in Computer Science

All highly cited Computer Science papers

Questions about this paper

How do I cite this paper in APA style?

Liu, T., Lyu, F., Li, S., Tang, C., Deng, Y., Zhou, C., & Wu, H. (2026). FedPKD + : A Prototype-Based Framework for Efficient and Flexible Heterogeneous Federated Learning. IEEE Transactions on Mobile Computing, 25(11), 20056–20073. https://doi.org/10.1109/tmc.2026.3710770

How many times has this paper been cited?

0 times, according to OpenAlex on 19 November 2026. Google Scholar and Scopus count different sets of sources, so their numbers differ.

Is this paper free to read?

Yes. It is open access at IEEE Transactions on Mobile Computing: sign in and open the full text from this page.

Who wrote this paper?

Tong Liu, Feng Lyu, Shucheng Li, Cheng Tang, Yongheng Deng, Conghao Zhou, Huaqing Wu.

What is the DOI of this paper?

10.1109/tmc.2026.3710770, permanently at https://doi.org/10.1109/tmc.2026.3710770. Sign in and open it from this page.

Bibliographic data from OpenAlex (CC0), last checked 19 November 2026. ResearchGuru does not host this paper; once you are signed in, the links open it at the publisher or the repository that does.