Deep Learning
A subset of machine learning using neural networks with many layers to learn complex hierarchical representations.
Definition
Deep learning uses artificial neural networks with multiple hidden layers to automatically learn hierarchical representations of data. Where traditional ML required hand-engineered features, deep learning learns features directly from raw data — pixels, words, audio waveforms.
The "deep" refers to the many layers of transformations. Each layer learns increasingly abstract representations: in image recognition, early layers detect edges, middle layers detect shapes, and later layers detect objects. This hierarchical feature learning drives performance on complex tasks.
Modern deep learning is powered by: (1) large datasets (ImageNet, Common Crawl), (2) GPU computing enabling parallel matrix operations, (3) architectural innovations (CNNs, RNNs, Transformers), and (4) regularisation techniques (dropout, batch normalisation). The 2012 AlexNet breakthrough ignited the current AI boom.
Examples
- Image classification
- GPT-4
- AlphaGo
- Voice recognition
Want a deeper dive?
Read our full explainer with use cases, how-it-works, and FAQs.
Deep Learning concept guideCompanies using this
Related Terms
Neural Network
A computing system of interconnected nodes inspired by biological brains, trained to recognise patterns.
Machine Learning (ML)
A subset of AI where systems learn patterns from data rather than following explicitly programmed rules.
Transformer
A neural network architecture using self-attention to process sequences in parallel — the foundation of all modern LLMs.
Convolutional Neural Network (CNN)
A neural network architecture using convolutional filters to extract spatial features from images or sequences.
Backpropagation
The algorithm for computing gradients in neural networks by propagating errors backwards through layers.