This particular algorithm is fascinating not just because it outperforms all other known neural networks on the given tasks, but because the same algorithm performs well on more than just one game. That is: it isn't restricted to strong performance on just one game, and its input actions can be extended at will (they are theoretically not just limited to the buttons on the Atari console). The algorithm is fairly simple, and the learning curve stabilizes already before 100 training epochs – which is quite impressive for the algorithm.
No wonder why Google is buying DeepMind Technologies, the company behind this feat. It'll be interesting to see what Google does with them, or rather what they end up doing for Google…
/via +Russ Abbott
Neural Net Learns Breakout Then Thrashes Human Gamers
The end is nigh. Humans have lost another key battle in the war against computer domination
This post has been reshared 6 times on Google+
View this post on Google+
Thank you, +Shaker Cherukuri I will peruse when I get home from work tonight.
+Dirk Reul Lot of info here:
https://plus.google.com/+ShakerCherukuri/posts/TqobaXULx7a
Interesting, in what direction do you think this would be going in that sector, +Shaker Cherukuri ?
+Dirk Reul +Sophie Wrobel possible energy sector applications I think. They don't need this for search.
There are, of course, the obvious candidates:
1. Google is going to push their automated driving systems further, using advanced AI, high quality robotics and sensor tech.
2. Google is trying to create interactive robots using the skills from Boston Dynamics, AI creation and sensor tech. Which would make a lot of sense, especially if combined with this little piece http://www.theguardian.com/technology/2013/sep/27/google-biggest-algorithm-change-hummingbird
So overall Google seems to be pushing into AI and robotics creation, possibly hoping for a feedback loop of both technologies that will enable growth for both.
3. Terminator and Skynet 😉
+Dirk Reul Good question. Would you care to speculate?
+Nikolai Varankine Much more interesting is what Google Scholar has to say about them: http://scholar.google.de/scholar?q=deepmind.com
Some of the stuff they've published on:
– Word similarity matching
– Universal intelligence
– Efficient policy gradient?
I fully agree, +Sophie Wrobel I wonder though why the push into artificial intelligence and robots including sensor data is being performed.
+Dirk Reul they're diversifying… to swallow up the whole vertical supply chain and be your one-hand-to give-it-all master supplier for the Internet of Things!
http://deepmind.com/ ? Very ascetic site 🙂
Humans employ 1-10 training epochs (just a speculation based on personal observations). That is likely a target. Having army of smart electronic "employees" turns business into higher profits.
This is quite fascinating. Now let's add the acquisition of Boston Dynamics and Nest and things get interesting.