By Patrick Lambe
For those of us who come into knowledge organisation (and knowledge management for that matter) via the taxonomy route, we are used to thinking of the field as a text-dominated field. However, there are exciting leapfrog developments in the artificial intelligence/ deep learning space as well. Image recognition (and matching), speech recognition, statistical methods of language translation (check out the new version of the Google Translate mobile app), and advances in the hardware required for this kind of computing and storage power, are all promising rapid shifts in being able to connect audio, video and image content with text-based content. One of our sessions at IKO will look at ways of building rich knowledgebases out of high resolution image details. Add facial recognition and auto-classification to that mix and you get instant propagation of insight into large historical image collections (for example). Here's a nice short piece on some of 2014's breakthroughs and promises in deep learning from Tom Simonite.
RSS Feed