{
  "id": "2009.09172",
  "version": "v1",
  "published": "2020-09-19T06:08:07.000Z",
  "updated": "2020-09-19T06:08:07.000Z",
  "title": "Few-shot learning using pre-training and shots, enriched by pre-trained samples",
  "authors": [
    "Detlef Schmicker"
  ],
  "categories": [
    "cs.LG",
    "cs.CV"
  ],
  "abstract": "We use the EMNIST dataset of handwritten digits to test a simple approach for few-shot learning. A fully connected neural network is pre-trained with a subset of the 10 digits and used for few-shot learning with untrained digits. Two basic ideas are introduced: during few-shot learning the learning of the first layer is disabled, and for every shot a previously unknown digit is used together with four previously trained digits for the gradient descend, until a predefined threshold condition is fulfilled. This way we reach about 90% accuracy after 10 shots.",
  "revisions": [
    {
      "version": "v1",
      "updated": "2020-09-19T06:08:07.000Z"
    }
  ],
  "analyses": {
    "keywords": [
      "few-shot learning",
      "pre-trained samples",
      "simple approach",
      "fully connected neural network",
      "pre-training"
    ],
    "note": {
      "typesetting": "TeX",
      "pages": 0,
      "language": "en",
      "license": "arXiv",
      "status": "editable"
    }
  }
}