4 ms·
This is an illuminating definition of deep learning from the end of the article: The goal of supervised deep learning is to solve almost any problem of the for
by grey-area 11y ago
This is an illuminating definition of deep learning from the end of the article:
The goal of supervised deep learning is to solve almost any problem of the form “map X to Y”. X can include images, speech, or text, and Y can include categories or even sentences.
Interesting to think about the implications when such systems become commonplace.
- tlrobinson 11y agoI know what they mean, but you could say the goal of any program is to "map X to Y" where X are the inputs and Y are the desired outputs :)
- grey-area 11y agoIt's a bit more precise than that because of the examples given, but also the interesting bit is that it does it with only general guidance from the programmer, on a set of x which is not previously known. Up to now algorithms have encoded domain knowledge, but that's about to change if deep learning lives up to its promise.
- zardo 11y agoThat's kind of the point. If you can solve the general case, you can solve all the instances too.
- chubot 11y agoOne thing I've heard talked about is to use deep learning to re-learn arbitrary programs from their inputs and outputs... Then you can throw away the program and just run the neural network. I'm sort of a skeptic, so I think that would just make the programs massively less efficient (usually), more incorrect, and harder to maintain, but I guess it's a research direction. Maybe programmers need to be put out of a job :)
- alphydan 11y ago> arbitrary programs It would be really fun to see a web app built on those principles! User clicks on /login/ link -> neural network {"hum ... I wonder what she's after ... let's send her a login form"} -> response. Trained on millions of MechanicalTurkers ... optimizing the function "which user can you keep engaged the longest". :)
- braderhart 11y agoI'm more interested to think about what happens when these structured systems become autonomous from unstructured data.
- grey-area 11y agoCan you elaborate, I'm not sure I understand?
- andreyk 11y agoSupervised learning is function approximation given examples of inputs and outputs, plain and simple. So yes, of course a function is anything that maps X to Y. Deep learning just does that with lots of computation steps.
- grey-area 11y agoPerhaps only illuminating if you don't know much about the subject already :)