4 ms·
I disagree that neural networks are necessarily mysterious. In most cases they can be straightforwardly understood as brute-force function approximation (approx
by mjn 10y ago
I disagree that neural networks are necessarily mysterious. In most cases they can be straightforwardly understood as brute-force function approximation (approximation here meaning mostly interpolation, in some cases capable of a limited degree of extrapolation). You take many samples of a function (either because you have a huge data set, or because you can actually sample with a simulator), and incrementally modify a large network to slowly converge it towards an accurate approximation of the function. The advent of very large data sets and GPU training made this possible, but the tech is pretty well understood, at least at the level of a 5-mile-high view.
- jules 10y agoOf course, but compared to how chess algorithms work it remains far more mysterious how exactly that produces an english description of a grid of pixel values.
- throwaway729 10y agoProbably only because every CS undergrad or self-trained person sees at least a few search algorithms. So although they probably couldn't write a performant chess-playing program, they get the general idea. I expect in 20 years undergraduate curricula evolve to cover NNs, and then those courses will have their salient bits adopted into mass market expository books/blog posts/code examples/etc. At that point someone like you will be saying something like you said about whatever the new hotness is.
- cosmolev 10y agoSo they take grid of pixels and produce an array of characters which may be interpreted by some humans (aka bags of chemicals).