“I’m not saying I want to forget deep learning... But we need to be able to extend it to do things like reasoning, learning causality, and exploring the world .” - Yoshua Bengio, not unlike what I have been saying since 2012 in The New Yorker.https://www.technologyreview.com/s/612434/one-of-the-fathers-of-ai-is-worried-about-its-future/ …
Part of the problem with this is that people actually waver on what they mean. The central force of my arguments for 20 years is that we need to add symbol-manipulation. Is that really what @ylecun or Yoshua have said? Hinton has ridiculed it, certainly. https://sites.google.com/site/krr2015/home/schedule …