AI Glossary
This glossary of artificial intelligence terms and concepts is designed to help you navigate the world of industrial AI, machine learning, and beyond.
An oversampling technique that is used to handle imbalanced datasets. It adapts the density distribution of the data by creating more synthetic data in the region of the feature space where the density of minority class examples is low, and fewer or none where the density is high.
In reinforcement learning, an agent is an entity that interacts with an environment, observes its current state, and takes actions to maximize some notion of cumulative reward.
A set of instructions designed to perform a specific task or solve a particular problem.
AI that is capable of performing any intellectual task that a human can.
The ability of machines to perform tasks that would normally require human intelligence.
A comprehensive approach to maintaining and improving the performance of physical assets, such as machinery, equipment, and infrastructure.
The ability of a system or machine to operate independently, without human intervention or control.
A statistical method for updating beliefs or probabilities based on new evidence or data.
Extremely large datasets that can be analyzed to reveal patterns, trends, and associations.
A measure of the amount of greenhouse gases (GHGs) emitted as a result of human activities, expressed in terms of the amount of carbon dioxide (CO2) that would have the same global warming potential over a specified time period.
A series of interconnected steps or processes that an AI system undergoes to arrive at a conclusion or answer.
An AI-powered program that simulates conversation with human users.
In machine learning, classification refers to a type of supervised machine learning technique that involves predicting a categorical or discrete output variable based on one or more input variables. The goal of classification is to build a model that can accurately classify new data points based on the patterns learned from the training data.
A control policy that uses feedback from the system being controlled to adjust or modify the control inputs in real-time.
A branch of computer science and artificial intelligence that focuses on the design and analysis of algorithms and models for machine learning.
A research laboratory at the Massachusetts Institute of Technology (MIT) dedicated to advancing the fields of computer science and artificial intelligence.
The ability of computers to interpret and analyze images and video.
A software system that helps organizations manage and track maintenance activities for their assets and equipment. CMMS is used in industries such as manufacturing, utilities, facilities management, defense, and transportation.
An AI model developed by OpenAI that is designed to understand and interpret visual concepts in a manner similar to how humans do, by connecting images and textual descriptions.
In general, a control is something that regulates or directs the behavior or operation of a system or process. In the context of engineering and technology, control refers specifically to the management or regulation of a system or process to achieve a desired outcome or to optimize performance.
The process of finding the best possible control policy for a given system, in order to achieve a desired outcome or to optimize performance.
The strategies or methods used to control or manage a system or process. In the context of engineering, control policies are used to manage or regulate a system or process to achieve a desired outcome or to optimize performance.
A framework created by the U.S. Department of Defense (DoD) to assess and enhance the cybersecurity posture of companies that do business with the DoD.
A strategy that allows significantly increasing the diversity of data available for training models, without actually collecting new data. This provides a way to add variability and improve the ability of models to generalize, leading to improved model performance.
The process of discovering patterns and insights in large datasets.
A series of processes or systems that are used to collect, process, and transform data from various sources and move it to a destination where it can be used for analysis or other applications.
An interdisciplinary field that involves the use of statistical, computational, and other quantitative methods to extract insights and knowledge from data.
Specific guidelines established to ensure the quality, accuracy, and consistency of the data entering a system.
A popular type of supervised learning algorithm used for both classification and regression tasks. They are based on a tree-like model of decisions and their possible consequences.
A type of machine learning that uses artificial neural networks to learn from data and make predictions.
A class of generative models that gradually adds random noise to a data sample (such as an image) until the original sample is completely transformed into noise. The model then reverses the process, starting from noise and learning to remove it to generate a new data sample.
A technique used to detect and diagnose faults in oil-insulated electrical equipment, such as transformers and circuit breakers.
A computerized control system used to monitor and control industrial processes, such as those in HVAC, power plants, chemical plants, and manufacturing facilities.
A project delivery method used in the construction industry. EPC contracts are commonly used in the construction of large-scale infrastructure projects, such as power plants, oil and gas facilities, and transportation systems.
A business management system that integrates various processes and functions (i.e., finance, human resources, manufacturing) across an organization into a single, centralized system.
AI systems that mimic the decision-making abilities of a human expert in a specific field.
The process of estimating values for input variables that lie outside the range of observed data points. It involves making predictions or forecasting for values that are beyond the scope of the training data.
A machine learning approach where a model is trained across multiple decentralized devices or servers each holding local data samples, without exchanging them.
A probabilistic model commonly used for regression and classification tasks. It is a non-parametric, flexible approach to modeling complex functions that can be used for tasks such as interpolation, extrapolation, and uncertainty quantification.
A category of artificial intelligence technologies that are designed to create new and original content, including text, images, videos, music, speech, and other types of media.
A type of artificial intelligence system where two neural networks, called the generator and the discriminator, are trained together in a sort of competition. The generator creates fake data that looks like the real data it's been trained on, while the discriminator tries to tell apart this fake data from the real data.
A type of statistical model that is designed to generate new data samples. These models learn the underlying distribution of a given dataset and can create new data points that resemble the original dataset.
A computer-based tool for storing, analyzing, and visualizing geospatial data. GIS allows users to capture, manage, and display data related to geographic locations, such as land use, topography, and population.
Systems used to provide indoor comfort and air quality in residential, commercial, and industrial buildings.
A practical, problem-solving approach that may not be perfect or optimal but is sufficient for reaching an immediate, short-term goal or approximation.
A statistical model used to model temporal data with an underlying hidden structure. HMM is a type of probabilistic graphical model that allows for the analysis and modeling of sequences of data, such as speech signals, biological sequences, or financial data.
A machine learning approach where the model learns continuously, updating its knowledge as new data comes in.
The use of Internet of Things (IoT) technologies and concepts in industrial settings, such as manufacturing, logistics, and energy production.
The process by which an AI system applies its trained model to new data to make predictions, draw conclusions, or take actions.
A professional responsible for designing, developing, and maintaining systems that control and monitor various processes and equipment in industrial settings. These systems may involve the use of sensors, actuators, controllers, and other devices to automate and optimize processes.
A class of control systems that utilize artificial intelligence techniques to manage and control complex systems.
A network of physical devices connected to the internet, which can be monitored and controlled remotely using AI.
In machine learning, interpolation refers to the process of estimating values for points within a range of observed data points.
A type of artificial intelligence model designed to understand, interpret, generate, and respond to human language. ChatGPT is an example of a large language model.
A metric used to evaluate the cost of producing energy from a particular source over its lifetime, expressed as the cost of producing one unit of energy (e.g. one kilowatt-hour or one megawatt-hour).
An important metric for logistics and supply chain management, LRT refers to the amount of time it takes for a logistics system to respond to a request for goods or services.
A type of recurrent neural network (RNN) architecture used for its effectiveness in processing and predicting sequences of data, making them ideal for tasks involving time series, natural language processing, and more.
A subset of AI that involves training algorithms to make predictions or decisions based on data.
A set of best practices and tools used to streamline and automate the process of deploying, managing, and monitoring machine learning models in production environments.
A metric used to measure the amount of time it takes to complete a maintenance task or cycle, from the moment a request is made until the task is completed.
A reliability metric that is used to estimate the expected time between the start of operation of a system or component and the occurrence of its first failure.
A mathematical framework or algorithm that is designed to learn from data and make predictions or decisions based on that data.
The process of automatically producing human-like text from data or structured information.
The ability of computers to understand and analyze human language.
A type of algorithm modeled after the structure of the human brain, used in deep learning.
A framework for generating photorealistic 3D models from a set of 2D images.
In artificial intelligence, particularly in neural networks, a node (also referred to as a neuron) is a fundamental processing element. Each node receives input, performs a computation on this input, and then passes its output to the next layer of nodes or as a final output.
In the context of Artificial Intelligence (AI) and machine learning, "noise" refers to unwanted or irrelevant information or distortions in data.
A control policy that does not use feedback from the system being controlled. Instead, it relies on a predetermined set of inputs or commands to achieve a desired output.
The specific state or condition of an operating system, machine, or device at a particular point in time.
The activities and processes involved in managing and maintaining a physical asset, such as a building, a machine, or an infrastructure system.
A common problem in machine learning and artificial intelligence, where a model performs well on the training data but poorly on unseen data (like validation or test data).
A type of computer simulation that relies on the laws of physics to predict or analyze real-world phenomena. Unlike data-driven models, which rely on analyzing large amounts of data to make predictions, physics-based models use established physical laws to predict outcomes.
In reinforcement learning, a policy refers to the strategy or rule that an agent uses to determine its actions in a given state.
In machine learning, prediction refers to the process of using a trained model to make predictions or forecasts about future or unseen data.
The process of using data, statistical algorithms, and machine learning techniques to identify the likelihood of future outcomes based on historical data.
A proactive maintenance strategy used in various industries to anticipate and prevent equipment failures before they occur, thereby reducing equipment downtime and costs.
A type of digital computer that is designed to control and automate industrial processes, machinery, and equipment.
A type of computing that uses quantum-mechanical phenomena, such as superposition and entanglement, to perform operations on data.
A technology that uses radio waves to identify and track objects, people, or animals.
A type of artificial neural network designed to recognize patterns in sequences of data, such as time series data, natural language, or audio.
A type of supervised machine learning technique used for predicting continuous numerical output values based on one or more input variables.
A type of ML in which an agent learns to make decisions by receiving rewards or punishments and using that information to alter its decision-making processes.
Used in predictive maintenance, RUL is the estimated time period that a piece of equipment or system can continue to operate before it reaches the end of its useful life.
Enables remote operators and managers to monitor and control equipment and processes from a centralized location, reducing the need for on-site inspections and maintenance.
The design, construction, and operation of robots to perform tasks.
A maintenance strategy in which equipment or assets are operated until they fail, without any preventive or proactive maintenance.
A type of machine learning approach where a combination of labeled and unlabeled data is used to train a model.
The process of using NLP to analyze and understand the emotions and opinions expressed in text.
A type of control system that is used in many industrial and engineering applications to regulate a process variable, such as temperature, pressure, or flow rate.
A computer-based system used to monitor and control industrial processes and infrastructure, such as power generation, water treatment, and oil and gas pipelines.
A machine learning technique used to handle class imbalance in a dataset. It's often used in scenarios where the number of instances of one class far outweighs the instances of the other class(es).
A reliability indicator used to measure the average outage duration of an electrical power distribution system.
A performance metric used to measure the reliability of an electric power distribution system, based on how many outages a customer experiences in a certain time period.
Data that varies over time, often in a continuous or sequential manner, i.e. weather patterns or stock market prices.
The duration of time between a specific starting point and a particular event of interest. The event can be any significant occurrence or milestone, such as the failure of a machine, the occurrence of a disease, or the completion of a project.
The set of data that is used to train a machine learning model, with the goal of teaching the machine learning algorithm to identify patterns and relationships in the data that can be used to make accurate predictions or classifications for new, unseen data.
A machine learning technique where a model developed for one task is reused as the starting point for a model on a second task. It is a popular approach in deep learning because it can train deep neural networks with comparatively little data.
Developed by Alan Turing, a test of a machine's ability to exhibit intelligent behavior equivalent to, or indistinguishable from, that of a human.
A global standard for measuring time that is based on the rotation of the Earth. It is used to synchronize clocks and timekeeping systems around the world, allowing for consistent time references across different time zones and regions. UTC is widely used in various fields, including telecommunications, aviation, finance, and computer systems.