Knowledge MaxxingDeep learning
Deep learning is a form of machine learning that uses neural networks with many stacked layers to learn increasingly abstract patterns directly from raw data.
Prerequisite: Machine learning
What matters
Each added layer lets the network combine simpler patterns from earlier layers into more complex ones, which is why deep networks can learn directly from raw pixels or text instead of requiring hand-engineered features.
Example
A deep network for photo tagging can learn to recognize edges in early layers, shapes in middle layers, and whole objects like 'dog' in later layers, without a human specifying any of those steps.
Common misconception
More layers do not automatically make a model better; a deep network with too little data or tuning can perform worse than a simpler model.
Recall question
What does stacking more layers let a deep network do that a shallow one cannot?