4 ms·

It works, and it's a marvel of engineering, but it's not "artificial intelligence." Is sound like you don't like complex algorithms written by humans. But you

by bh42222 15y ago

It works, and it's a marvel of engineering, but it's not "artificial intelligence." Is sound like you don't like complex algorithms written by humans. But you do like a big bucket of "neural network"? How is "The magical black box works somehow!" better than "We know exactly how this white box we built works."?

mattmanser 15y ago
Posing your question differently, you get a completely different understanding of why the black box is much more alluring. Which is better, something that can learn or something that does exactly what we told it to?
- kd1220 15y ago
  Now I pose another question to you. What is "learning?" When you feed a bunch of input to an ANN, isn't it doing exactly what it's told? The ANN itself does not seek out knowledge, it only adjusts weights according to an algorithm depending on inputs. The allure to having a fixed, visible algorithm is that you are aware of many of the limitations. Having a black box obscures those limitations, but I won't deny it's alluring in other ways.
- bh42222 15y ago
  Or how about: Which is better, something which can learn exactly as we programmed it to, or something which learns by some unknown way in which the big 'ol bucket of bytes orders itself? Or, if you had a magic learning machine, wouldn't you want to know exactly what its learning mechanisms are?
brandonb 15y ago
The black box can adapt to new circumstances. A lot of the algorithms used today in speech recognition were established on much smaller data sets (thousands of hours of speech); the tradeoffs made then may not apply when you have 1000x the data. The more automatic the algorithm, the more it can change. Existing speech recognizers aren't really a "white box". You can't look at a Gaussian Mixture Model and understand what it's doing. The more you automate the whole training process, the more.

by bh42222 15y ago

mattmanser 15y ago

Posing your question differently, you get a completely different understanding of why the black box is much more alluring. Which is better, something that can learn or something that does exactly what we told it to?

kd1220 15y ago
Now I pose another question to you. What is "learning?" When you feed a bunch of input to an ANN, isn't it doing exactly what it's told? The ANN itself does not seek out knowledge, it only adjusts weights according to an algorithm depending on inputs. The allure to having a fixed, visible algorithm is that you are aware of many of the limitations. Having a black box obscures those limitations, but I won't deny it's alluring in other ways.
bh42222 15y ago
Or how about: Which is better, something which can learn exactly as we programmed it to, or something which learns by some unknown way in which the big 'ol bucket of bytes orders itself? Or, if you had a magic learning machine, wouldn't you want to know exactly what its learning mechanisms are?

brandonb 15y ago

The black box can adapt to new circumstances. A lot of the algorithms used today in speech recognition were established on much smaller data sets (thousands of hours of speech); the tradeoffs made then may not apply when you have 1000x the data. The more automatic the algorithm, the more it can change. Existing speech recognizers aren't really a "white box". You can't look at a Gaussian Mixture Model and understand what it's doing. The more you automate the whole training process, the more.