{
  "id": "2409.01985",
  "version": "v3",
  "published": "2024-09-03T15:26:51.000Z",
  "updated": "2025-01-22T21:52:30.000Z",
  "title": "UNSURE: self-supervised learning with Unknown Noise level and Stein's Unbiased Risk Estimate",
  "authors": [
    "Julián Tachella",
    "Mike Davies",
    "Laurent Jacques"
  ],
  "journal": "ICLR 2025",
  "categories": [
    "stat.ML",
    "cs.LG",
    "eess.SP"
  ],
  "abstract": "Recently, many self-supervised learning methods for image reconstruction have been proposed that can learn from noisy data alone, bypassing the need for ground-truth references. Most existing methods cluster around two classes: i) Stein's Unbiased Risk Estimate (SURE) and similar approaches that assume full knowledge of the distribution, and ii) Noise2Self and similar cross-validation methods that require very mild knowledge about the noise distribution. The first class of methods tends to be impractical, as the noise level is often unknown in real-world applications, and the second class is often suboptimal compared to supervised learning. In this paper, we provide a theoretical framework that characterizes this expressivity-robustness trade-off and propose a new approach based on SURE, but unlike the standard SURE, does not require knowledge about the noise level. Throughout a series of experiments, we show that the proposed estimator outperforms other existing self-supervised methods on various imaging inverse problems.",
  "revisions": [
    {
      "version": "v3",
      "updated": "2025-01-22T21:52:30.000Z"
    }
  ],
  "analyses": {
    "subjects": [
      "68U10",
      "I.4.5",
      "G.3"
    ],
    "keywords": [
      "steins unbiased risk estimate",
      "unknown noise level",
      "self-supervised learning",
      "assume full knowledge",
      "similar cross-validation methods"
    ],
    "tags": [
      "journal article"
    ],
    "note": {
      "typesetting": "TeX",
      "pages": 0,
      "language": "en",
      "license": "arXiv",
      "status": "editable"
    }
  }
}