To follow up on yesterday's post on deep learning, this fantastic feature article on Geoff Hinton - one of the pioneers of deep learning. Deep learning works by layering the neural nets to process at successive levels of detail, with feedback loops to reinforce the successful patterns. His breakthrough in machine learning was to allow the neural networks to learn patterns first without labelling each layer - only when the machine had learnt enough would labels be applied to the layers to assign meaning for human consumption. Actually this is exactly how good taxonomy development should begin - gather the data first, figure out meaningful patterns and clusters, assign concept labels later. It's usually disastrous to go in with a preconceived label-set and try to assign whatever you have to it. As for early practical applications? Google can now read YouTube videos and learn - by itself - how to recognise cats. Lesson? Learn first, categorise later.

"Nobody is saying that this system has exceeded the human ability to classify photos; indeed, if a human hired to write captions performed at the level of this neural net, the newbie wouldn’t last until lunchtime. But it did shockingly, shockingly well for a machine. Some of the dead-on hits included “a group of young people playing a game of frisbee,” “a person riding a motorcycle on a dirt road,” and “a herd of elephants walking across a dry grass field.” Considering that the system “learned” on its own concepts like a Frisbee, road, and herd of elephants, that’s pretty impressive." Hat tip to Andrew McAfee for this.

By Patrick Lambe

In this video I explain how we were inspired to put together the IKO Conference.