Deep Learning
A type of machine learning that uses neural networks with many layers to learn complex patterns directly from large amounts of raw data.
Deep learning is the most powerful branch of machine learning. It uses neural networks made of many stacked layers — the “depth” that gives the technique its name.
Each layer learns progressively more abstract features. For an image, early layers might detect edges, middle layers detect shapes, and later layers recognise whole objects. Because the network discovers these features automatically, deep learning excels at messy, high-dimensional data like images, audio, and language.
Deep learning needs large datasets and significant computing power, which is why it only became dominant in the 2010s. It now powers voice assistants, image recognition, translation, and the generative AI behind modern chatbots. Read the full explainer in our guide to deep learning.