News for hackers. Stuff that matters.
Re:Deep Learning, the Curse of Dimensionality, and Autoencoders by galapago on Wednesday April 01, 2015 @07:00AM (#9302761)
> Autoencoders are an extremely exciting new approach to unsupervised learning

They are not new.

[ Parent | on HN ]
Re:Deep Learning, the Curse of Dimensionality, and Autoencoders by jostmey on Wednesday April 01, 2015 @09:22AM (#9303296)
Yeah, but some of methods that make the technique successful are less than ten years old.
[ Parent | on HN ]
Re:Deep Learning, the Curse of Dimensionality, and Autoencoders by idunning on Wednesday April 01, 2015 @02:14PM (#9305625)
Original content is at http://nikhilbuduma.com/2015/03/10/the-curse-of-dimensionali...
[ Parent | on HN ]
Re:Deep Learning, the Curse of Dimensionality, and Autoencoders by mercurial on Wednesday April 01, 2015 @07:32AM (#9302866)
Seems to me that you would seriously reduce the size of the problem by first running a face-recognition algorithm, and then only analyzing the face...
[ Parent | on HN ]
Re:Deep Learning, the Curse of Dimensionality, and Autoencoders by chaosfactor on Wednesday April 01, 2015 @11:36AM (#9304407)
You've created a nice little chicken and egg problem for yourself.
[ Parent | on HN ]
Re:Deep Learning, the Curse of Dimensionality, and Autoencoders by im3w1l on Wednesday April 01, 2015 @12:02PM (#9304608)
I think he means first detect that there is a face at all, and where in the image it is. When you know that, do further processing.
[ Parent | on HN ]
Re:Deep Learning, the Curse of Dimensionality, and Autoencoders by mercurial on Wednesday April 01, 2015 @03:49PM (#9306323)
That's the idea.
[ Parent | on HN ]
Re:Deep Learning, the Curse of Dimensionality, and Autoencoders by bayesianhorse on Wednesday April 01, 2015 @08:40AM (#9303096)
Sounds a bit like the age old method of producing medical professionals: Make them read and reproduce large amounts of medical facts.
[ Parent | on HN ]