Deep learning refers to neural networks that are much deeper than the three to four layers people were using before. It's interesting because, in order to make it practical at all, you have to use a pretraining step which involves pretending that it's a totally different thing called a restricted Boltzmann machine. That pretraining, if you're lucky, acts as a form of feature extraction and can save feature engineering time.
The initial successes seemed almost magical and everyone leapt on it. It turned out that those successes were very specialized and, while it's a really useful tool, it's not the panacea people thought it might be.
For further actions, you may consider blocking this person and/or reporting abuse
We're a place where coders share, stay up-to-date and grow their careers.
Deep learning refers to neural networks that are much deeper than the three to four layers people were using before. It's interesting because, in order to make it practical at all, you have to use a pretraining step which involves pretending that it's a totally different thing called a restricted Boltzmann machine. That pretraining, if you're lucky, acts as a form of feature extraction and can save feature engineering time.
The initial successes seemed almost magical and everyone leapt on it. It turned out that those successes were very specialized and, while it's a really useful tool, it's not the panacea people thought it might be.