This was the strangest moment of COSYNE. Deepmind rep gets on stage visibly frustrated, says networks with biologically plausible credit assignment (non-backprop) perform terribly. Indeed seems like a challenge issued to the community.https://twitter.com/neurosutras/status/970365749129859077 …
-
Show this thread
-
Replying to @butterflyarson
I will say that the presence of data in that talk that showed AI outperforming humans on the ImageNet task makes me think that maybe brains are just not optimal computers or something
2 replies 2 retweets 7 likes -
Replying to @JChartove @butterflyarson
Shock “computer outperforms human on task computer is built for more than humans are” news now...
1 reply 1 retweet 6 likes -
Yep. It's actually nothing like we categorise things. http://karpathy.github.io/2014/09/02/what-i-learned-from-competing-against-a-convnet-on-imagenet/ …
1 reply 1 retweet 8 likes -
Replying to @o_guest @PsychScientists and
I'm hoping to beat my calculator at long division one of these days!
2 replies 0 retweets 1 like
To be fair, you can win on some metrics. You can go to more decimal points than your calculator. 
-
-
-
This Tweet is unavailable
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.
