Machine Learning Glossary

AI Fairness

AI Fairness is a concept in the field of artificial intelligence sciences that concerns the creation and application of algorithms that ensure unbiased decisions. It aims to promote equality and prevent any form of discrimination (like age, race, or gender discrimination) which might be inadvertently incorporated in AI system's decision-making process.

AI Observability

AI Observability refers to the capacity to understand and interpret the processes, decisions, and actions of an AI system. It emphasizes making the system's internal workings transparent and decipherable, allowing for a clearer understanding of how inputs are transformed into outputs.

Activation Function

An activation function is a mathematical function applied to a neuron's output, determining the neuron's output or "activation." It introduces non-linearity into the network, enabling the model to learn from error and make complex predictions.

Active Learning

Active learning is a special case of machine learning in which a learning algorithm can interactively query a user to label new data points with the desired outputs.

Adversarial Machine Learning

Adversarial machine learning is the study of the attacks on machine learning algorithms, and of the defenses against such attacks, typically by introducing specially crafted input data (known as adversarial examples) to cause incorrect model predictions.

Annotation

Annotation refers to the process of adding metadata, labels, or additional information to raw data. This metadata or information can be used to train supervised machine learning models, provide insights, or make the data more accessible and understandable.

Anomaly Detection

Anomaly detection, also known as outlier detection, refers to the identification of patterns in a dataset that do not conform to expected behavior. These non-conforming patterns are termed anomalies or outliers, and they often provide critical and actionable information in various application domains.

Artificial Neural Network

An Artificial Neural Network (ANN) is a computational model inspired by the way human brains work. It's designed to recognize patterns and consists of interconnected nodes or "neurons".

Augmentation

Data augmentation refers to techniques used to artificially increase the size of a dataset by applying various transformations on the original data. Having a larger training dataset can lead to better generalization.

AutoML

AutoML, short for Automated Machine Learning, refers to the automated process of end-to-end development of machine learning models. This includes tasks like data preprocessing, feature engineering, model selection, hyperparameter tuning, and model evaluation, all with minimal human intervention.

Autoencoder

An autoencoder is a type of artificial neural network used for unsupervised learning of efficient codings or representations of data, typically for the purpose of dimensionality reduction or feature extraction.

Autoregressive Model

An Autoregressive Model (AR Model) is a representation of a type of random process; as such, it is used to describe certain time-varying processes in nature, economics, etc. In this model, the output variable relies on its own previous values and a stochastic term. It is usually employed for forecasting in time series analysis, as it takes a specific period of historical data and uses it to predict future values.

Average Precision

Average Precision (AP) is a metric commonly used in the context of binary classification and information retrieval to summarize the Precision-Recall curve. It provides a single number that characterizes the quality of the ranked retrieval results for a given class or category, especially in tasks like object detection.

BERT

BERT, which stands for Bidirectional Encoder Representations from Transformers, is a natural language processing (NLP) model developed by Google in 2018 for tasks such as question answering, language understanding, and sentiment analysis. This deep learning model has significantly improved the understanding of the context of words in a sentence, and goes beyond examining words in isolation.

BLEU

BLEU (Bilingual Evaluation Understudy) is an algorithm used for evaluating the quality of text that has been translated from one language to another. It is a metric for evaluating a generated sentence to a reference sentence.

Backpropagation

Backpropagation is a popular method used in training artificial neural networks. It involves training a system to refine its predictions or outputs by propagating the error from the output backwards through each layer of the network to modify weights as well as biases. It's a subset of supervised learning algorithms, based on the gradient descent optimization algorithm.

Bagging (Bootstrap Aggregating)

Bagging, an acronym for Bootstrap Aggregating, is an ensemble machine learning algorithm. The general idea of bagging is to create several subsets of data from the original dataset, with replacement, and train a model for each subset. The final prediction is calculated by averaging the predictions (in case of regression problem) or by voting (in case of a classification problem) from each model.

Baseline Models

Baseline models refer to simple or rudimentary models utilized in machine learning. They serve as a point of comparison for more complex models developed during the machine learning tasks. They are out-of-the-box models that require minimal effort and are easy to implement but can set a minimum level of performance that more sophisticated models will need to exceed.

Batch Normalization

Batch Normalization is a technique used in training deep neural networks to standardize the inputs to a layer for each mini-batch. This process helps to stabilize the learning process and dramatically reduces the number of training steps required to converge.

Bayesian Inference

Bayesian inference is a statistical method which gives a framework to update probabilities of a hypothesis based on evidence or additional data. This method is rooted in the principle of Bayesian probability, a theory that describes the probability of an event, based on prior knowledge of conditions that might be related to the event.

Bias-variance Tradeoff

Bias-variance tradeoff is a fundamental concept in machine learning that refers to the delicate balance between bias (underfitting) and variance (overfitting) in predictive models.

Binary Classification

Binary Classification is a type of machine learning algorithm that categorizes data into one of two distinct classes. The output is a binary outcome, that is, either 'Yes' or 'No', 'True' or 'False', 'Positive' or 'Negative', etc. This makes it easier to make effective decisions and predictions based on past data.

Binominal Distribution

The binomial distribution is a probability distribution that describes the number of successes in a fixed number of independent Bernoulli trials with the same probability of success. It is used when there are exactly two mutually exclusive outcomes of a trial often labelled as "success" and "failure". The distribution is defined by two parameters: the probability of success in a single trial and the number of trials.

Black Box Model

A black box model refers to a system or process where the internal workings or details are unknown or not understood, and it's classified or evaluated only based on its inputs and outputs. The term is frequently used in a variety of fields like finance, computing, and machine learning. It originates from the notion of a 'black box', an object that you can't see inside, but you know what goes in and what comes out.

Bounding Box

A bounding box refers to the smallest rectangular or cuboidal box that can completely enclose a given set of points or a complex object. The box is defined by minimum and maximum coordinates along each dimension. In the world of AI and computer vision, a bounding box is often used for object detection, where it's used to define the location and size of the identified objects within an image or a video frame.

CI/CD

CI/CD refers to the application of Continuous Integration and Continuous Deployment techniques in the realm of Machine Learning. These practices play a critical role in increasing productivity and efficiency of teams and organizations in managing ML projects.