Nice video, however instead of riding the hype train of arxiv, could we wait until peer review analyzes the paper?
If someone other than Hinton presented a YADLA (Yet Another Deep Learning Architecture) that does not achieve state of the art level of performance in the basic datasets, it would not be very well received.
If someone other than Hinton presented a YADLA (Yet Another Deep Learning Architecture) that does not achieve state of the art level of performance in the basic datasets, it would not be very well received.