Artificial Intelligence

See the Alan Turing Institute's Data Science and AI Glossary for these and other definitions.

Algorithm
A sequence of rules that a computer uses to complete a task. An algorithm takes an input (e.g. a dataset) and generates an output (e.g. a pattern that it has found in the data). Algorithms underpin the technology that makes our lives tick, from smartphones and social media to sat nav and online dating, and they are increasingly being used to make predictions and support decisions in areas as diverse as healthcare, employment, insurance, and law.

Artificial Intelligence
The design and study of machines that can perform tasks that would previously have required human (or other biological) brainpower to accomplish. AI is a broad field that incorporates many different aspects of intelligence, such as reasoning, making decisions, learning from mistakes, communicating, solving problems, and moving around the physical world. AI was founded as an academic discipline in the mid-1950s, and is now found in myriad everyday applications, including virtual assistants, search engines, navigation apps and online banking.

Data
Any information that has been collected for analysis or reference. Data can take the form of numbers and statistics, text, symbols, or multimedia such as images, videos, sounds and maps. Data that has been collected but not yet processed, cleaned or analysed is known as ‘raw’ or ‘primary’ data.

Deep Learning
A form of machine learning that uses computational structures known as ‘neural networks’ to automatically recognise patterns in data and provide a suitable output, such as a prediction or evidence for a decision. Deep learning neural networks are loosely inspired by the way neurons in animal brains are organised, being composed of multiple layers of simple computational units (‘neurons’), and they are suited to complex learning tasks such as picking out features in images and speech. Deep learning thus forms the basis of the voice control in our phones and smart speakers, and enables driverless cars to identify pedestrians and stop signs. See also ‘neural network’.

Generative AI
(Turing) An artificial intelligence system that generates text, images, audio, video or other media in response to user prompts. It uses machine learning techniques to create new data that has similar characteristics to the data it was trained on (see ‘generative adversarial network’), resulting in outputs that are often indistinguishable from human-created media (see ‘deepfake’).

Large Language Model
 A type of foundation model that is trained on a vast amount of textual data in order to carry out language-related tasks. Large language models power the new generation of chatbots, and can generate text that is indistinguishable from human-written text. They are part of a broader field of research called natural language processing, and are typically much simpler in design than smaller, more traditional language models.

Machine Learning (ML)
A field of artificial intelligence involving computer algorithms that can ‘learn’ by finding patterns in sample data. The algorithms then typically apply these findings to new data to make predictions or provide other useful outputs, such as translating text or guiding a robot in a new setting. Medicine is one area of promise: machine learning algorithms can identify tumours in scans, for example, which doctors might have missed.

Natural Language Processing (NLP)
A field of artificial intelligence that uses computer algorithms to analyse or synthesise human speech and text. The algorithms look for linguistic patterns in how sentences and paragraphs are constructed, and how the words, context and structure work together to create meaning. Applications include speech-to-text converters, chatbots, speech recognition, automatic translation, and sentiment analysis (identifying the mood of a piece of text).

Neural Network
An artificial intelligence system inspired by the biological brain, consisting of a large set of simple, interconnected computational units (‘neurons’), with data passing between them as between neurons in the brain. Neural networks can have hundreds of layers of these neurons, with each layer playing a role in solving the problem. They perform well in complex tasks such as face and voice recognition. See also ‘deep learning’.
 feedforward manner during training and inference.

Supervised, Semisupervised, Unsupervised Learning
In artificial intelligence (AI), learning methods are categorized into unsupervised, supervised, and semi-supervised learning, each involving the human programmer's role in distinct ways.

Unsupervised learning tasks entail algorithms analyzing input data without explicit instructions from human programmers on desired outputs, allowing the algorithm to discover patterns or structures within the data autonomously.

Supervised learning requires human programmers to provide input-output pairs during training, where each input is associated with a known output, enabling the algorithm to learn from labeled examples and map inputs to corresponding outputs effectively.

Semi-supervised learning utilizes both labeled and unlabeled data during training, with human programmers providing a smaller set of labeled data alongside a larger pool of unlabeled data. This approach aims to enhance model performance by leveraging the labeled data while also allowing the algorithm to learn from the broader unlabeled dataset autonomously.