marysia-winkels-equivariance-in-cnns-how-generalising-the-weight-sharing-pydata-berlin-2019.json

{
  "description": "Speaker: Marysia Winkels\n\nTrack:PyData\nIn this talk, we will explore how the weight-sharing property of the convolutional layer can be generalised to achieve equivariance towards transformations beyond just translation, how to implement this, and the results on real-world data.\n\nRecorded at the PyConDE & PyData Berlin 2019 conference.\nhttps://pycon.de\n\nMore details at the conference page: https://de.pycon.org/program/RAS8UK\nTwitter:  https://twitter.com/pydataberlin\nTwitter:  https://twitter.com/pyconde",
  "language": "eng",
  "recorded": "2019-10-10",
  "related_urls": [
    {
      "label": "Conference Schedule",
      "url": "https://de.pycon.org/program/RAS8UK"
    }
  ],
  "speakers": [
    "Marysia Winkels"
  ],
  "thumbnail_url": "https://i.ytimg.com/vi/t7U-Z5a7oWw/hqdefault.jpg",
  "title": "Equivariance in CNNs: how generalising the weight-sharing...",
  "videos": [
    {
      "type": "youtube",
      "url": "https://www.youtube.com/watch?v=t7U-Z5a7oWw"
    }
  ]
}