H

Definition: A technique designed for solving a problem more quickly when classic methods are too slow, or for finding an approximate solution when classic methods fail to find any exact solution. This approach is often used in decision-making, learning, and problem-solving.

Definition: A parameter whose value is used to control the learning process and is set before the learning process begins. Unlike model parameters, hyperparameters are not derived via training.

Definition: In a neural network, a layer of neurons that is neither an input nor output layer. Hidden layers transform inputs into something that the output layer can use.

Definition: A method of cluster analysis which seeks to build a hierarchy of clusters. It is either agglomerative (starting with individual points and aggregating them into clusters) or divisive (starting with a single cluster and dividing it into smaller clusters).

Definition: A mathematical optimisation technique which belongs to the family of local search. It is an iterative algorithm that starts with an arbitrary solution to a problem and then attempts to find a better solution by incrementally changing a single element of the solution.

Definition: An algorithm for detecting simple shapes such as circles, lines, and ellipses in an image, typically used in image analysis and digital image processing.

Definition: Specialised hardware used to speed up specific aspects of a computational task. In AI, accelerators like GPUs and TPUs are used to speed up the training and inference phases of deep learning models.

Definition: The process of converting an input (or ‘message’) into a fixed-size string of bytes. The output, typically a ‘digest’, represents concisely the original data. This is used in various applications in computer science including indexing and retrieving items in databases, load balancing, and more.

Definition: A form of encryption that allows computation on ciphertexts, generating an encrypted result which, when decrypted, matches the result of operations performed on the plaintext. This is useful in data privacy for cloud computing and outsourcing of scientific computations.

Definition: A hypothesis in neuroscience that proposes an increase in synaptic efficacy arises from the presynaptic cell’s repeated and persistent stimulation of the postsynaptic cell. It is often summarised as “Cells that fire together, wire together.”

Definition: Technology that recreates the sense of touch by applying forces, vibrations, or motions to the user, enhancing the realism of virtual simulations.

Definition: A form of recurrent artificial neural network popularised by John Hopfield in 1982, which serves as content-addressable memory systems with binary threshold nodes.

Definition: A model of interaction where a human is involved in the loop of a machine-learning algorithm, providing inputs, feedback, or decisions to influence the outcome.

Definition: In AI, a hybrid model combines characteristics of both neural networks and algorithmic approaches, typically to leverage the strengths of each in solving a problem.

Definition: In machine learning, particularly in support vector machines, a hyperplane is a decision boundary that separates a set of objects having different class memberships.

Definition: Data that comprises a mixture of text, numbers, dates, and sometimes images and videos, which can present additional challenges in data processing and analysis.

Definition: A graphical representation of the distribution of numerical data. It is an estimate of the probability distribution of a continuous variable and was first introduced by Karl Pearson.

Definition: A sequence of rescaled “square-shaped” functions which together form a wavelet family or basis. Wavelets are used in signal processing for tasks such as image compression and noise reduction.

Definition: A measure of the minimum number of substitutions required to change one string into the other, or the number of errors that transformed one string into the other. This concept is used in various branches of computer science, including error detection and correction.

Definition: A method used in user interface design to evaluate a user interface. In machine learning, heuristic methods might be used to optimise model parameters.

Definition: In mathematical optimisation, this square matrix of second-order partial derivatives of a scalar-valued function describes the local curvature of a function of many variables. The Hessian matrix is used in machine learning to help find the minimum (or maximum) of a function.

Definition: Data with many attributes or features, which can complicate modelling techniques such as regression analysis due to the increased complexity and computational burden.