{
  "abstract": "Statistical heterogeneity limits federated learning when a single global classifier cannot represent client-specific label distributions. In this work, we propose Personalized Federated Knowledge Distillation with Head Adaptation (pFedKDH), which aggregates only the shared backbone, keeps persistent client-specific heads, and uses a recalibrated global head as a teacher during local training. Across MNIST, Fashion-MNIST, CIFAR10, and CIFAR100 under class-wise Dirichlet partitions, pFedKDH obtains the best accuracy in most settings, with accuracy gaps up to 37.67\\% over the weakest baseline and consistently low standard deviation across repetitions. Component-wise diagnostics and convergence results support the role of persistent heads and distillation-guided local optimization under label-skewed data.",
  "abstract_shown": true,
  "affiliations": [],
  "article_type": null,
  "arxiv": "2609.17284",
  "author_total": 3,
  "authors": [
    {
      "name": "Polycarpo Souza Neto"
    },
    {
      "name": "José Mairton Barros da Silva Júnior"
    },
    {
      "name": "Charles Casimiro Cavalcante"
    }
  ],
  "bibtex": "https://research.draconitic.com/paper/c32frjqtq95zg.bib",
  "cited_by": 0,
  "doi": "10.48550/arxiv.2609.17284",
  "fulltext_url": null,
  "id": "c32frjqtq95zg",
  "issns": [],
  "landing_url": "https://arxiv.org/abs/2609.17284",
  "language": null,
  "licence": "LicenseRef-arXiv-nonexclusive-1.0",
  "licence_label": "arXiv licence",
  "licence_source": "article",
  "licence_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
  "metadata_licence": "CC0-1.0",
  "native_subjects": [
    [
      "arxiv",
      "cs.LG"
    ],
    [
      "arxiv",
      "stat.ML"
    ],
    [
      "arxiv",
      "stat.OT"
    ],
    [
      "openalex",
      "1702|Privacy-Preserving Technologies in Data"
    ]
  ],
  "openalex": "W7213430384",
  "pdf_url": "https://arxiv.org/pdf/2609.17284",
  "pmcid": null,
  "pmid": null,
  "published": "2026-09-15",
  "publisher": null,
  "ris": "https://research.draconitic.com/paper/c32frjqtq95zg.ris",
  "sources": [
    {
      "first_seen": "2026-10-30T12:21:51Z",
      "native_id": "2609.17284",
      "source": "arxiv",
      "url": "https://arxiv.org/abs/2609.17284"
    }
  ],
  "subjects": [
    {
      "field": "mathematics",
      "origin": "arxiv:stat.ML",
      "subfield": "statistics"
    }
  ],
  "title": "Personalized Federated Learning through Global Knowledge Distillation and Local Head Adaptation",
  "title_text": "Personalized Federated Learning through Global Knowledge Distillation and Local Head Adaptation",
  "type": "preprint",
  "updated_at": "2026-10-30T12:21:51Z",
  "url": "https://research.draconitic.com/paper/c32frjqtq95zg",
  "venue": "arXiv",
  "versions": []
}