{
  "id": "1812.06080",
  "version": "v1",
  "published": "2018-12-14T18:59:03.000Z",
  "updated": "2018-12-14T18:59:03.000Z",
  "title": "Online gradient-based mixtures for transfer modulation in meta-learning",
  "authors": [
    "Ghassen Jerfel",
    "Erin Grant",
    "Thomas L. Griffiths",
    "Katherine Heller"
  ],
  "categories": [
    "cs.LG",
    "stat.ML"
  ],
  "abstract": "Learning-to-learn or meta-learning leverages data-driven inductive bias to increase the efficiency of learning on a novel task. This approach encounters difficulty when transfer is not mutually beneficial, for instance, when tasks are sufficiently dissimilar or change over time. Here, we use the connection between gradient-based meta-learning and hierarchical Bayes (Grant et al., 2018) to propose a mixture of hierarchical Bayesian models over the parameters of an arbitrary function approximator such as a neural network. Generalizing the model-agnostic meta-learning (MAML) algorithm (Finn et al., 2017), we present a stochastic expectation maximization procedure to jointly estimate parameter initializations for gradient descent as well as a latent assignment of tasks to initializations. This approach better captures the diversity of training tasks as opposed to consolidating inductive biases into a single set of hyperparameters. Our experiments demonstrate better generalization performance on the standard miniImageNet benchmark for 1-shot classification. We further derive a novel and scalable non-parametric variant of our method that captures the evolution of a task distribution over time as demonstrated on a set of few-shot regression tasks.",
  "revisions": [
    {
      "version": "v1",
      "updated": "2018-12-14T18:59:03.000Z"
    }
  ],
  "analyses": {
    "keywords": [
      "online gradient-based mixtures",
      "transfer modulation",
      "leverages data-driven inductive bias",
      "experiments demonstrate better generalization performance",
      "meta-learning"
    ],
    "note": {
      "typesetting": "TeX",
      "pages": 0,
      "language": "en",
      "license": "arXiv",
      "status": "editable"
    }
  }
}