3 ms·
It seems like that kind of transition could happen badly. I've done pure math to the graduate level and I've a modest amount of programming. But even with both
by joe_the_user 10y ago
It seems like that kind of transition could happen badly.
I've done pure math to the graduate level and I've a modest amount of programming. But even with both these backgrounds, I've bitten hard by efforts to do applied math. It seems that applied math is very the domain of people who both know algorithms and know how to "squint correctly" at a given problem and see whether X algorithm is appropriate. And that seems to be a very necessary skill. Where the validity of one approach stops and another begins is fairly opaque until one is an expert.
The thing is that neural network appear to be fairly similar. They're somewhat generic but tuning again requires high level expertise and tutelage.
The worst case scenario is that deep learning solves a broad but still limited class of problem, expanding somewhat the nebulous area cover by applied math. Then all the expertise and credibility goes there and gradually, the nebulous area covered by deep learning stops expanding and progress actually becomes harder since there's less real knowledge, just general intuition about a single, quite general approach.
- giardini 10y ago* ...people who both know algorithms and know how to "squint correctly" at a given problem ...* * ... tuning again requires high level expertise and tutelage.* Appears to be more luck (or "serendipity", if one is generous) that is required! As new algorithms crop up, everyone moves to them to redo older problems. That continues until the next new thing pops up. Pedro Domingo's book, The Master Algorithm", describes this somewhat: http://www.amazon.com/Master-Algorithm-Ultimate-Learning-Machine/dp/0465065708?ie=UTF8&keywords=the%20master%20algorithm&qid=1464207298&ref_=sr_1_1&s=books&sr=1-1 http://www.amazon.com/Master-Algorithm-Ultimate-Learning-Mac...