Deep learning.

Deep learning trains multilayer neural networks to transform inputs into useful predictions or outputs. “Deep” refers to the layers of computation, not to humanlike understanding.

Each layer transforms information for the next. In an image model, early layers can detect edges, while later layers combine features into more complex patterns. Training adjusts the network’s parameters to improve a chosen objective, allowing useful features to be learned from examples.

Deep learning supports image recognition, speech processing, and generative models. Large models can require substantial data and compute; using a pretrained model can reduce the work needed for a new task. Performance still depends on data quality and evaluation on unfamiliar examples.

Not all machine learning is deep learning, and not all deep learning generates content. “Deeper” also does not automatically mean more accurate.