{
  "id": 3077,
  "url": "https://arxiv.org/abs/2606.28123v1",
  "title": "Dangerous Liaisons of Convex Learning and Non-Affine Aggregation",
  "summary": "Last-iterate convergence and generalization guarantees in first-order convex learning hinge on the monotonicity of the update operator. While linear averaging preserves the monotonicity of gradient updates, this property is often violated when gradients are aggregated non-affinely, as in modern pipelines enforcing constraints like adaptivity, privacy, robustness or fairness. Whether it is possible to design non-affine aggregation rules that maintain monotonicity has remained an open question. We",
  "authors": "Thomas Boudou, Batiste Le Bars, Nirupam Gupta, Aurélien Bellet",
  "category": "research",
  "topics": "bias-fairness,privacy-surveillance",
  "orgs": null,
  "regions": null,
  "published_at": "2026-06-26T14:24:11.000Z",
  "fetched_at": "2026-07-14T16:11:46.979Z",
  "source_slug": "x-arxiv-fairness-query",
  "source_name": "arXiv fairness query",
  "source_homepage": "https://arxiv.org",
  "ethics_ai_record_url": "https://ethics.ai/record/3077",
  "original_url": "https://arxiv.org/abs/2606.28123v1",
  "evidence_status": "source-only",
  "attribution": "via ethics.ai"
}