{
  "id": "1811.07579",
  "version": "v2",
  "published": "2018-11-19T09:45:20.000Z",
  "updated": "2019-09-05T11:05:25.000Z",
  "title": "Deep Active Learning with a Neural Architecture Search",
  "authors": [
    "Yonatan Geifman",
    "Ran El-Yaniv"
  ],
  "comment": "Accepted to NeurIPS 2019",
  "categories": [
    "cs.LG",
    "stat.ML"
  ],
  "abstract": "We consider active learning of deep neural networks. Most active learning works in this context have focused on studying effective querying mechanisms and assumed that an appropriate network architecture is a priori known for the problem at hand. We challenge this assumption and propose a novel active strategy whereby the learning algorithm searches for effective architectures on the fly, while actively learning. We apply our strategy using three known querying techniques (softmax response, MC-dropout, and coresets) and show that the proposed approach overwhelmingly outperforms active learning using fixed architectures.",
  "revisions": [
    {
      "version": "v2",
      "updated": "2019-09-05T11:05:25.000Z"
    }
  ],
  "analyses": {
    "keywords": [
      "neural architecture search",
      "deep active learning",
      "overwhelmingly outperforms active learning",
      "deep neural networks",
      "appropriate network architecture"
    ],
    "note": {
      "typesetting": "TeX",
      "pages": 0,
      "language": "en",
      "license": "arXiv",
      "status": "editable"
    }
  }
}