What Is Machine Learning?

Machine learning is the part of artificial intelligence concerned with systems that improve task performance by learning patterns from data. Instead of depending only on fixed instructions, a machine learning system is trained on examples so it can generalize to new inputs.

Short answer

Machine learning is a method for building AI systems by training models on data.

The model does not memorize every example exactly. It learns statistical patterns that help it make useful predictions, classifications, rankings, or generations when it sees new inputs later.

Why it matters

Machine learning is the practical core of modern AI.

Many widely used systems depend on it, including recommendation engines, fraud detection, ranking systems, speech recognition, and language models. If you understand machine learning, you understand why data, task definition, and evaluation matter so much in AI.

Machine learning also gives structure to the earlier idea in How AI Systems Learn From Data. That earlier article explains the learning pattern. This article names the field built around that pattern.

How it works

In machine learning, developers usually define:

• a task
• a dataset
• a model
• an objective for improvement

During training, the model processes examples and updates its parameters so the objective improves. The learning signal depends on the problem type. In supervised learning, the model is trained against known targets. In unsupervised settings, it finds structure without explicit labels. In reinforcement learning, feedback comes from rewards rather than direct answers.

The result is a model that can apply learned patterns to new cases during inference.

Where it fits

Machine learning sits inside artificial intelligence as one of the most important modern approaches. It is broader than deep learning, because deep learning is one family of machine learning methods rather than the whole field.

That means the relationship is:

• AI is the umbrella field
• machine learning is a major method inside AI
• deep learning is a major method inside machine learning

This layering matters because later parts of the hub build on it. Large language models, retrieval systems, and many agent systems ultimately depend on machine learning foundations.

Common questions