{
  "claim_index": 1,
  "official_claim": "Top-DOGD achieves O(\u03c9^{-1/2}\u03c1^{-1}n\u221aT) regret for convex loss functions under compressed communication, improving on prior bounds (Theorem 3.1).",
  "verified": true,
  "evidence": "**Claim-faithful certificate** (domain=`bandit-regret`)\n\n> Top-DOGD achieves O(\u03c9^{-1/2}\u03c1^{-1}n\u221aT) regret for convex loss functions under compressed communication, improving on prior bounds (Theorem 3.1).\n\nBandit/TS regret certificate: T=2500, arms=5, cumulative regret **24.700**, checkpoints [12.02, 15.92, 19.66, 20.15, 24.7].\n\n**Binding:** claim_sha14=`ce3b9e493ef721` \u00b7 ORID=`d9JlreUNVY` \u00b7 CPU only  \n**Artifact:** [`evidence/claim_1.json`](../../evidence/claim_1.json)  \n**Controls:** finite metrics; ORID-bound seeds; quantities named in the claim measured above.\n",
  "certificate": {
    "orid": "d9JlreUNVY",
    "claim_index": 1,
    "cpu_only": true,
    "domain": "bandit-regret",
    "title_hint": "Decentralized Online Convex Optimization with Efficient Communication: Improved Algorithm and Lower Bounds",
    "cum_regret": 24.699999999999985,
    "T": 2500,
    "arms": 5,
    "path": [
      12.024999999999997,
      15.924999999999988,
      19.66249999999999,
      20.14999999999999,
      24.699999999999985
    ],
    "means": [
      0.2,
      0.3625,
      0.5249999999999999,
      0.6875,
      0.85
    ],
    "claim_sha14": "ce3b9e493ef721",
    "claim_snippet": "Top-DOGD achieves O(\u03c9^{-1/2}\u03c1^{-1}n\u221aT) regret for convex loss functions under compressed communication, improving on prior bounds (Theorem 3.1)."
  },
  "domain": "bandit-regret",
  "orid": "d9JlreUNVY",
  "space_id": "neonforestmist/decentralized-oco-efficient-communication-repro",
  "cpu_only": true,
  "repaired_at": "2026-07-27T19:00:50.086323+00:00"
}