In plain English
Deep learning is a powerful kind of machine learning that stacks many layers of simple maths units. The more layers, the more complex the patterns it can learn, like recognising speech or writing text.
In practice
Most AI breakthroughs of the last decade, from voice assistants to ChatGPT, are deep learning. It needs large amounts of data and computing power, which is why most organisations use pre-trained models rather than training their own.
Under the hood
Deep learning trains multi-layer neural networks end to end with backpropagation and gradient descent. Depth lets a network learn hierarchical representations, for example edges, then shapes, then objects. Common architectures include convolutional networks, recurrent networks and transformers.
Example
"Deep learning made real-time speech-to-text accurate enough for meeting transcripts."