{
  "id": "2003.09671",
  "version": "v1",
  "published": "2020-03-21T14:43:45.000Z",
  "updated": "2020-03-21T14:43:45.000Z",
  "title": "On Information Plane Analyses of Neural Network Classifiers -- A Review",
  "authors": [
    "Bernhard C. Geiger"
  ],
  "comment": "8 pages; under review",
  "categories": [
    "cs.LG",
    "cs.CV",
    "cs.IT",
    "math.IT",
    "stat.ML"
  ],
  "abstract": "We review the current literature concerned with information plane analyses of neural network classifiers. While the underlying information bottleneck theory and the claim that information-theoretic compression is causally linked to generalization are plausible, empirical evidence was found to be both supporting and conflicting. We review this evidence together with a detailed analysis how the respective information quantities were estimated. Our analysis suggests that compression visualized in information planes is not information-theoretic, but is rather compatible with geometric compression of the activations.",
  "revisions": [
    {
      "version": "v1",
      "updated": "2020-03-21T14:43:45.000Z"
    }
  ],
  "analyses": {
    "keywords": [
      "neural network classifiers",
      "information plane analyses",
      "information bottleneck theory",
      "information-theoretic compression",
      "geometric compression"
    ],
    "note": {
      "typesetting": "TeX",
      "pages": 8,
      "language": "en",
      "license": "arXiv",
      "status": "editable"
    }
  }
}