Over the last 18 months, we have been collaborating with a research project called the Congruence Engine. One strand of the project has been experimenting with new ways of using our Historical Directories collection. Daniel Wilson from the project has written a blog post this week getting into the technical side of this.
The workshop I attended around Christmas was focused on testing computer vision methods – how to get a computer to ‘read’ images (rather than words). If you have had a go with Dall-E or Stable Diffusion recently, similar ideas are lying behind how those tools work. If you would like to find out more about this project, do get in touch.
(You can also read part one of Daniel’s reflections, which explains why directories are such a great source of information about the past).