{
  "id": "2004.14046",
  "version": "v1",
  "published": "2020-04-29T09:45:06.000Z",
  "updated": "2020-04-29T09:45:06.000Z",
  "title": "Reducing catastrophic forgetting with learning on synthetic data",
  "authors": [
    "Wojciech Masarczyk",
    "Ivona Tautkute"
  ],
  "categories": [
    "cs.LG",
    "stat.ML"
  ],
  "abstract": "Catastrophic forgetting is a problem caused by neural networks' inability to learn data in sequence. After learning two tasks in sequence, performance on the first one drops significantly. This is a serious disadvantage that prevents many deep learning applications to real-life problems where not all object classes are known beforehand; or change in data requires adjustments to the model. To reduce this problem we investigate the use of synthetic data, namely we answer a question: Is it possible to generate such data synthetically which learned in sequence does not result in catastrophic forgetting? We propose a method to generate such data in two-step optimisation process via meta-gradients. Our experimental results on Split-MNIST dataset show that training a model on such synthetic data in sequence does not result in catastrophic forgetting. We also show that our method of generating data is robust to different learning scenarios.",
  "revisions": [
    {
      "version": "v1",
      "updated": "2020-04-29T09:45:06.000Z"
    }
  ],
  "analyses": {
    "keywords": [
      "synthetic data",
      "reducing catastrophic forgetting",
      "two-step optimisation process",
      "real-life problems",
      "learn data"
    ],
    "note": {
      "typesetting": "TeX",
      "pages": 0,
      "language": "en",
      "license": "arXiv",
      "status": "editable"
    }
  }
}