{
  "id": "1812.03928",
  "version": "v1",
  "published": "2018-12-10T17:26:25.000Z",
  "updated": "2018-12-10T17:26:25.000Z",
  "title": "Learning Representations of Sets through Optimized Permutations",
  "authors": [
    "Yan Zhang",
    "Jonathon Hare",
    "Adam Prügel-Bennett"
  ],
  "categories": [
    "cs.LG",
    "cs.CV",
    "stat.ML"
  ],
  "abstract": "Representations of sets are challenging to learn because operations on sets should be permutation-invariant. To this end, we propose a Permutation-Optimisation module that learns how to permute a set end-to-end. The permuted set can be further processed to learn a permutation-invariant representation of that set, avoiding a bottleneck in traditional set models. We demonstrate our model's ability to learn permutations and set representations with either explicit or implicit supervision on four datasets, on which we achieve state-of-the-art results: number sorting, image mosaics, classification from image mosaics, and visual question answering.",
  "revisions": [
    {
      "version": "v1",
      "updated": "2018-12-10T17:26:25.000Z"
    }
  ],
  "analyses": {
    "keywords": [
      "learning representations",
      "optimized permutations",
      "image mosaics",
      "achieve state-of-the-art results",
      "traditional set models"
    ],
    "note": {
      "typesetting": "TeX",
      "pages": 0,
      "language": "en",
      "license": "arXiv",
      "status": "editable"
    }
  }
}