{
  "id": "2004.12289",
  "version": "v1",
  "published": "2020-04-26T05:15:36.000Z",
  "updated": "2020-04-26T05:15:36.000Z",
  "title": "Deep k-NN for Noisy Labels",
  "authors": [
    "Dara Bahri",
    "Heinrich Jiang",
    "Maya Gupta"
  ],
  "comment": "Full paper (including supplemental) can be found at https://github.com/dbahri/deepknn",
  "categories": [
    "cs.LG",
    "cs.AI",
    "stat.ML"
  ],
  "abstract": "Modern machine learning models are often trained on examples with noisy labels that hurt performance and are hard to identify. In this paper, we provide an empirical study showing that a simple $k$-nearest neighbor-based filtering approach on the logit layer of a preliminary model can remove mislabeled training data and produce more accurate models than many recently proposed methods. We also provide new statistical guarantees into its efficacy.",
  "revisions": [
    {
      "version": "v1",
      "updated": "2020-04-26T05:15:36.000Z"
    }
  ],
  "analyses": {
    "keywords": [
      "noisy labels",
      "deep k-nn",
      "modern machine learning models",
      "remove mislabeled training data",
      "nearest neighbor-based filtering approach"
    ],
    "tags": [
      "github project"
    ],
    "note": {
      "typesetting": "TeX",
      "pages": 0,
      "language": "en",
      "license": "arXiv",
      "status": "editable"
    }
  }
}