Machine learning

Definition

Machine learning is a subset of artificial intelligence that enables systems to learn patterns from data and improve performance without being explicitly programmed. Instead of relying on fixed rules, machine learning models adapt and make predictions or decisions based on new inputs.

It is widely applied in industries such as finance, healthcare, retail, and technology for tasks like fraud detection, recommendation systems, forecasting, and personalisation.

Advanced

At an advanced level, machine learning uses algorithms that fall into categories such as supervised learning, unsupervised learning, and reinforcement learning. Supervised learning trains models on labelled data, unsupervised learning discovers hidden patterns in unlabeled data, and reinforcement learning optimises actions through trial and error.

Deep learning, a specialised form of machine learning, uses neural networks with multiple layers to process large datasets and handle tasks like image recognition, speech processing, and natural language understanding. Advanced applications require significant computational power, often using GPUs or cloud infrastructure.

Why it matters

• Transforms raw data into actionable insights.
• Automates predictions and decision-making at scale.
• Enhances customer experiences through personalisation.
• Provides businesses with a competitive edge in innovation and efficiency.

Use cases

• Fraud detection in banking and payments.
• Personalised recommendations in e-commerce and streaming.
• Predictive maintenance in manufacturing.
• Diagnostic tools in healthcare based on medical imaging.

Metrics

• Model accuracy, precision, and recall.
• Training time and computational efficiency.
• Error rates such as false positives and false negatives.
• Business impact is measured by ROI and cost savings.

Issues

• Requires large amounts of high-quality data.
• Risk of bias and fairness concerns in model predictions.
• Interpretability challenges in complex models.
• Security vulnerabilities, such as adversarial attacks.

Example

A streaming platform uses machine learning to analyse viewing history and recommend shows. The system learns user preferences over time, improving engagement and increasing subscription retention.