{
  "id": 5977,
  "url": "https://arxiv.org/abs/2604.10849v1",
  "title": "Task2vec Readiness: Diagnostics for Federated Learning from Pre-Training Embeddings",
  "summary": "Federated learning (FL) performance is highly sensitive to heterogeneity across clients, yet practitioners lack reliable methods to anticipate how a federation will behave before training. We propose readiness indices, derived from Task2Vec embeddings, that quantifies the alignment of a federation prior to training and correlates with its eventual performance. Our approach computes unsupervised metrics -- such as cohesion, dispersion, and density -- directly from client embeddings. We evaluate t",
  "authors": "Cristiano Mafuz, Rodrigo Silva",
  "category": "research",
  "topics": "safety-alignment,healthcare",
  "orgs": null,
  "regions": null,
  "published_at": "2026-04-12T22:48:51.000Z",
  "fetched_at": "2026-07-14T16:32:11.181Z",
  "source_slug": "arxiv-ethics",
  "source_name": "arXiv",
  "source_homepage": "https://arxiv.org",
  "ethics_ai_record_url": "https://ethics.ai/record/5977",
  "original_url": "https://arxiv.org/abs/2604.10849v1",
  "evidence_status": "source-only",
  "attribution": "via ethics.ai"
}