7 ms·

I agree. The majority of DL layers are about feature engineering, not performing classification.

• wpietri 4y ago
  Could you say more about this? One of the things that interests me about nominal AI applications is the extent to which they're sort of a Mechanical Turk or what I've heard called Artificial Artificial Intelligence. By which I mean it's sold as computer magic, but most of the magic is actually humans sneaking in a lot of human judgement. That can come through humans directly massaging the output or through human-driven selection of results. But I've also been wondering to what extent natural human intelligence is getting put in at the lower layers of the system, like feature engineering.
  • jb_s 4y ago
    looks like AI, quacks like a bunch of linear equations
    • eru 4y ago
      Lots of linear functions with \x -> max(0, x) thrown in between. (That's literally what neural nets with relu as activation unit do.)
  • laichzeit0 4y ago
    This is how I see it: Statistical modeling: input -> feature extraction (manual) -> model selection (manual) -> output Machine learning: input -> feature extraction (manual) -> model selection (auto) -> output Deep learning: input -> feature extraction (auto) -> model selection (auto) -> output So take a DL image classifier. The convolution + pooling layers perform automatic feature extraction. Back to OPs point, why use something like DL when you've already engineered your features?
    • wpietri 4y ago
      Wow, what a good summary. Thanks!