Machine learning is easier than it looks

Re:Machine learning is easier than it looks by robrenaud on Wednesday November 20, 2013 @09:43PM (#6772742)

I can totally appreciate the value of having well defined and well understood definitions and notations that might require lots of background to understand. This allows for precise and concise discourse.

But there have been many published bounds in Machine Learning Theory that turned out to be completely vacuous. That is, they prove a bound on generalization error that is always greater than 1.

http://hunch.net/?p=181

To me, that definitely seems like complicated math for the sake of showing off and writing papers rather than writing for clarity and demonstrating something useful.

[ Parent | on HN ]

Re:Machine learning is easier than it looks by nico on Wednesday November 20, 2013 @10:32PM (#6772899)

I think some people just have a better grasp of the technical language that is commonly used for describing most of this stuff. The language definitely helps the people most closely involved, but it also leaves a lot of people out.

Being an abstraction, the language helps in developing the field further just by allowing to find new abstractions on top of the already existing ones. Very similar to what happens with programming languages in general.

[ Parent | on HN ]

Re:Machine learning is easier than it looks by zodiac on Saturday November 23, 2013 @09:10AM (#6785951)

It would be math for the sake of showing off only if the authors or referees knew before publishing that the bounds were vacuous. In fact, most mathematicians would probably be embarrassed if that happened to one of their results, and it's nothing to show off about.

[ Parent | on HN ]

Re:Machine learning is easier than it looks by kskz on Thursday November 21, 2013 @02:00PM (#6776608)

Not really. This is more of an example of a "failed experiment" than showing off. Research is hard. You never know if what you're doing will be useful, or just plain wrong.

[ Parent | on HN ]