Machine learning vs deep learning
Two terms people use interchangeably that are not the same. How they nest, and what actually changed with "deep".
The short version
- → Deep learning is a subset of machine learning, which is a subset of AI.
- → Machine learning means a program that improves from examples instead of hand-written rules.
- → Deep learning uses many-layered neural networks that learn their own features.
These terms nest inside each other. Artificial intelligence is the widest circle: any technique that makes machines act smart. Machine learning is a circle inside it. Deep learning is a smaller circle inside that.
The confusion comes from the fact that almost every headline-grabbing AI result of the last decade, including large language models, is deep learning. So the words get used as if they mean the same thing. They do not.
Machine learning: learning from examples
In traditional software a person writes explicit rules. In machine learning you instead show the program many examples and it finds the patterns itself. Spam filters, product recommendations, and credit scoring are classic machine learning. Much of it does not use neural networks at all.
Deep learning: many-layered networks
Deep learning is machine learning built on neural networks with many layers stacked together. The "deep" just refers to the number of layers. More layers let the model learn its own useful features from raw data, rather than needing a human to hand-engineer them.
The thing deep learning removed
Older machine learning often needed experts to decide which features of the data mattered, for example which pixels or which words to measure. Deep learning learns those features on its own, straight from raw input. That is why it scaled to messy data like images, audio, and free text.
Where LLMs sit
A large language model is deep learning: a very large, many-layered neural network (a transformer) trained on text. So it is deep learning, which is machine learning, which is AI. All three labels are technically correct for it.
An analogy
Think of nested boxes. AI is the big box. Inside it is a machine learning box. Inside that is a deep learning box. Inside that sits the language model. Every inner box is also everything the outer boxes are.
Go deeper
How a neural network works
By the end you will have computed a small neural network by hand and know exactly what a large one is doing more of. 17 min read →
Where Berges AI fits
The models Berges AI runs are all deep learning systems. Knowing the hierarchy mostly helps you read the field clearly: when a product says "AI-powered", it usually means one specific deep learning model doing one specific job.
Related concepts
How AI models are trained
The process that turns a random network into something useful: read a lot, adjust, repeat, then teach it to behave. Read →
What is an LLM?
The kind of AI behind chat assistants, explained without the jargon: what it is, how it produces text, and where it falls short. Read →
Things people ask
Is deep learning always better than machine learning?
No. Deep learning shines on large, messy data like images and text, but it needs a lot of data and compute. For smaller, structured problems, simpler machine learning is often faster, cheaper, and easier to interpret.
Is a neural network the same as deep learning?
A neural network with many layers is deep learning. A very shallow one is still a neural network but would not usually be called "deep". The depth is the point.
Where does generative AI fit?
Generative AI, the models that produce text or images, is a use of deep learning. It sits in the innermost circle alongside language models.