News for hackers. Stuff that matters.

A comment on Deep Learning’s Impact on Image Processing, Mathematics, and Humanity. Parent comment

Re:Deep Learning’s Impact on Image Processing, Mathematics, ... by zamalek on Sunday May 07, 2017 @06:16AM (#14284733)
If the internet were around when microwave ovens were invented, there would have been much the same outcry. Deep learning is excessively simple and elegant, I can explain it in person in under 5 minutes to anyone who has done high school algebra.
[ Parent | on HN ]
Re:Deep Learning’s Impact on Image Processing, Mathematics, ... by reader5000 on Sunday May 07, 2017 @07:18AM (#14284903)
Great. Explain why training converges in subexponential time despite the nonconvex cost function and does not overfit the data.
[ Parent | on HN ]
Re:Deep Learning’s Impact on Image Processing, Mathematics, ... by sprobertson on Sunday May 07, 2017 @08:35AM (#14285132)
As you get closer to the right answer, the changes you need to make are smaller and harder to discover.
[ Parent | on HN ]
Re:Deep Learning’s Impact on Image Processing, Mathematics, ... by zamalek on Sunday May 07, 2017 @08:02AM (#14285030)
You only need to know that if you are studying the field. A good explanation of deep learning is no more surprising to the general populace than general relativity is (spacetime isn't actually the surface of a balloon).
[ Parent | on HN ]
Re:Deep Learning’s Impact on Image Processing, Mathematics, ... by reader5000 on Sunday May 07, 2017 @10:03AM (#14285490)
That's not correct though, theres plenty of reasons why on an intuitive level neural networks should not work. Also, genetic algorithms are intuitively appealing but do not work well.
[ Parent | on HN ]
Re:Deep Learning’s Impact on Image Processing, Mathematics, ... by singham on Sunday May 07, 2017 @04:04PM (#14287303)
Genetic Algorithms are used in designing of antenna arrays.
[ Parent | on HN ]
Re:Deep Learning’s Impact on Image Processing, Mathematics, ... by jph00 on Sunday May 07, 2017 @02:27PM (#14286758)
Because in high dimensional space there are partial saddle points, not local minima. Intuitively obvious, and shown theoretically in the last 12 months.
[ Parent | on HN ]
Re:Deep Learning’s Impact on Image Processing, Mathematics, ... by reader5000 on Sunday May 07, 2017 @07:18PM (#14288239)
Are you referencing https://arxiv.org/abs/1412.0233?
[ Parent | on HN ]
Re:Deep Learning’s Impact on Image Processing, Mathematics, ... by soVeryTired on Monday May 08, 2017 @04:45AM (#14290113)
Shown theoretically under fairly artificial assumptions.
[ Parent | on HN ]