E

Definition: A machine learning technique where multiple models, such as classifiers or experts, are strategically generated and combined to solve a particular computational intelligence problem.

Definition: In the context of training an artificial neural network, an epoch is one complete presentation of the data set to be learned to a learning machine. Each epoch consists of one full training cycle on the training set.

Definition: A subset of evolutionary computation, these algorithms use mechanisms inspired by biological evolution, such as reproduction, mutation, recombination, and selection, to develop solutions to optimisation and search problems.

Definition: AI systems that emulate the decision-making ability of a human expert. They are designed to solve complex problems by reasoning through bodies of knowledge, represented mainly as if-then rules rather than through conventional procedural code.

Definition: A representation of data where elements of the same type (such as words or products) are mapped to points in the same space. This is often used in natural language processing to reduce the dimensionality of text data.

Definition: A form of regularisation used to avoid overfitting when training a learning algorithm iteratively. It involves stopping the training process if the performance on a validation dataset degrades.

Definition: Also known as Named Entity Recognition (NER), this is the process of identifying and classifying key elements from text into predefined categories, such as the names of persons, organisations, locations, expressions of times, quantities, monetary values, percentages, etc.

Definition: Refers to methods and techniques in the application of artificial intelligence technology such that the results of the solution can be understood by human experts.

Definition: A framework typically used in machine translation where one neural network encodes a source sentence and a second neural network decodes that information to translate it into another language.

Definition: A regularised regression method that linearly combines the L1 and L2 penalties of the lasso and ridge methods used in statistics.

Definition: In neural networks, the error gradient is a vector that points in the direction of the steepest increase in error. Gradient descent algorithms use this to update the weights of the network, minimising the error.

Definition: A distributed computing paradigm that brings computation and data storage closer to the location where it is needed, to improve response times and save bandwidth.

Definition: The physical equipment like smartphones, sensors, and other gadgets that collect or generate data at the outer boundary of a network. These devices often have the capability to perform some basic data processing tasks directly on them before sending the data elsewhere for further analysis.

Definition: Techniques that create multiple models and then combine them to produce improved results. Popular methods include boosting, bagging, and stacking.

Definition: In the context of AI, it’s a type of matrix factorisation that involves decomposing a matrix into its eigenvalues and eigenvectors. This is useful for various kinds of applications, including those that involve understanding the properties of graphs and networks.

Definition: A strategy used primarily in reinforcement learning where the agent chooses the best current action with high probability (1 – epsilon), and a random action with a small probability (epsilon), thus balancing the exploration-exploitation trade-off.

Definition: In AI, it refers to a category of memory that involves the ability to recall specific events from the past, similar to human memory. This is particularly studied in the context of developing lifelike robots and AI agents.

Definition: A common metric used in machine learning for calculating the distance between two points in Euclidean space. It serves as a basis for many clustering algorithms.

Definition: Refers to the process of automatically extracting structured information from unstructured and/or semi-structured machine-readable documents.

Definition: In biological and artificial neural networks, these neurons activate other neurons, sending them signals to increase their activity.

Definition: A rule of thumb technique for smoothing time series data using the exponential window function. In AI, it’s often used in forecasting models.

Definition: A technique used in machine learning where multiple predictions are averaged to improve the robustness and accuracy of predictive models.

Definition: Computer systems with a dedicated function within a larger mechanical or electrical system, often with real-time computing constraints. It is embedded as part of a complete device often including hardware and mechanical parts.

Definition: In the context of machine learning, entropy is often used as a measure of impurity or randomness when creating decision trees. Lower entropy means less impurity.

Definition: A type of genetic algorithm that uses mechanisms inspired by biological evolution such as mutation, selection, and reproduction to evolve programs or solutions to problems.