B

A method used in artificial neural networks to improve their accuracy by adjusting the weights of the neurons based on the error rate obtained in the previous run (learning from mistakes).

A type of statistical model that represents a set of variables and their conditional dependencies via a directed acyclic graph (DAG). It’s used for decision making and risk assessment.

In machine learning, bias is the tendency of an AI system to consistently make errors in judgment that favour one outcome over others. It can result from assumptions in the machine learning process.

Extremely large data sets that may be analysed computationally to reveal patterns, trends, and associations, especially relating to human behaviour and interactions.

A type of classification task where an AI model must decide between two choices, such as yes/no, true/false, or pass/fail.

The measurement and statistical analysis of people’s physical and behavioural characteristics, used primarily for identification and access control, such as fingerprints or face recognition technologies.

A software application that runs automated tasks (scripts) over the internet. Bots can vary in complexity from performing relatively simple tasks to mimicking human behavior in interactions.

A processing mode where a group of data is collected over time and then processed all at once. This is common in machine learning for tasks where real-time processing is not necessary.

A fundamental problem in supervised learning where a model needs to balance the bias (errors from erroneous assumptions) against variance (errors from sensitivity to small fluctuations in the training set).

An ensemble technique that creates a strong classifier from a number of weak classifiers by focusing on training instances that previous classifiers misclassified.

A concept where the decision-making capacities of agents (including AI systems) are limited by the information they have, the cognitive limitations of their minds, and the finite amount of time they have to make decisions.

A system of recording information in a way that makes it difficult or impossible to change, hack, or cheat the system. It’s a digital ledger of transactions that is duplicated and distributed across the entire network of computer systems on the blockchain.

AI technologies and systems designed to handle a wide range of tasks and environments, rather than being specialised for one specific function.

Automated programs designed to simulate conversation with human users, especially over the internet, often used in customer service and information acquisition.

An algorithm design paradigm for solving combinatorial optimisation problems. It involves systematising the enumeration of candidate solutions by dividing them into smaller, more manageable sub-problems.

A hidden method for bypassing normal authentication or securing unauthorised remote access to a computer, while remaining undetected.

A method in robotics and artificial intelligence where the actions of a human are mimicked or learned by a machine using observed data from the human’s actions.

An algorithm used in Bayesian networks, graphical models, and error-correcting codes, for calculating the distributions of variables and making inferences about them.

In AI, benchmarks are standard tests used to measure and compare the performance of different AI systems or models on specific tasks.

In AI development, a detailed plan or map of the architecture and functionality of an AI system or application.

Areas or aspects where an AI system lacks the ability to see, understand, or process information, which can lead to errors or oversight in operations.

A direct communication pathway between an enhanced or wired brain and an external device, often used in various types of medical and technological applications.

A simple problem-solving technique that systematically enumerates all possible candidates for the solution and checks whether each candidate satisfies the problem’s statement.