Deep Learning

Working knowledgeCore AI Concepts

Deep learning is a subset of machine learning that uses neural networks with many layers. It is the technique driving modern breakthroughs in computer vision and language processing. The layers let a system learn useful features directly from raw data, rather than requiring a person to specify in advance what to look for.

In practice

Deep learning is why modern AI needs so much data and compute, and why its decisions are hard to explain: the features it learns are not readable by a person. Higher accuracy bought at the price of lower interpretability is a trade a regulated business should make deliberately, use case by use case.

Not sure where your organisation stands?

Take the free AI-readiness diagnostic.

Start the diagnostic