• Google AI program DeepMind learns human navigation skills
    6 replies, posted
https://www.theguardian.com/technology/2018/may/09/googles-ai-program-deepmind-learns-human-navigation-skills We're all gonna die
This is such a vague article Scientists noticed that when they trained the AI to move through a landscape, it spontaneously developed electrical activity akin to that seen in the specialised brain cells that underpin human navigational skills. How can software develop "electrical activity"? What's that even supposed to mean? It went on to beat experienced players at a game that involved racing through rooms to find a prize after being dropped into the virtual environment at a random location. I was under the impression that even simple scripted AI is better at maze puzzles than humans because humans are comparatively awful at those. I think the takeaway here is that the AI developed a sort of emulation of brain cells for spatial cognition, which is impressive but the article doesn't actually shine any light in the technicalities of that
The architecture used is already emulating brain cells, I think what they mean is that the pattern of signals going between these neurons was similar to how signals propagate through the parts of our brains that handle spatial reasoning.
Which is........ duh, literally the basic unit and largest revolution in machine learning, the perception, was specifically modeled after brain neurons in how they connect with each otger
as it turns out, humans really are in fact at their peak, and robots are just chasing to catch up.
How long until ai's go into rooms, forget why they went in there, and leave again.
Evolutionary convergence maybe. Faced with same problem as animal brains once had, AI developed exact functionality to solve it without having common ancestor (not even being organic in the first place lol).
Sorry, you need to Log In to post a reply to this thread.