AI Terms Glossary: AI Terms in 2024
A
AGI
Artificial General Intelligence (AGI) is a type of AI that can understand, learn, and apply knowledge in any domain, similar to human intelligence.
AI
Artificial Intelligence (AI) refers to the simulation of human intelligence in machines.
ANN
Artificial Neural Network, a framework for many different machine learning algorithms to process complex data inputs.
API
Application Programming Interface (API) is a set of rules and protocols for building and interacting with software applications.
AR
Augmented Reality (AR) is a technology that overlays digital information onto the real world, typically through a device such as a smartphone or AR glasses.
Active Learning
A machine learning approach where the model identifies the data it is uncertain about or considers most informative, and queries a user or oracle to label it.
Adversarial Attack
A technique used to fool models through malicious input.
Alexa
Alexa is a virtual assistant developed by Amazon, known for its integration with the Amazon Echo smart speaker.
Algorithm
A set of rules or instructions given to an AI or ML system to help it learn from data.
Autoencoder
A type of artificial neural network used to learn efficient codings of unlabeled data, typically for the purpose of dimensionality reduction.
Autonomous Vehicle
An Autonomous Vehicle (AV) is a vehicle capable of navigating and operating without human input.
B
BERT
Bidirectional Encoder Representations from Transformers (BERT) is a natural language processing model developed by Google.
Backpropagation
A method used in artificial neural networks to calculate the error contribution of each neuron after a batch of data is processed.
Batch Normalization
A technique to improve the speed, performance, and stability of artificial neural networks by normalizing the inputs of each layer.
Bayesian Networks
A type of probabilistic graphical model that uses Bayesian inference for probability computations.
Belief Networks
Graphical models that represent the probabilistic relationships among a set of variables.
Bias
A systematic error in predictions, due to assumptions in the learning algorithm.
Big Data
Extremely large data sets that may be analyzed computationally to reveal patterns, trends, and associations.
Blockchain
Blockchain is a decentralized and distributed digital ledger technology used to record transactions across multiple computers.
C
CNN
Convolutional Neural Network, a deep learning algorithm which can take in an input image, assign importance to various aspects/objects in the image.
CV
Computer Vision (CV) is the field of AI that focuses on enabling computers to interpret and understand visual information.
CaaS
Conversational AI as a Service (CaaS) refers to cloud-based platforms that provide conversational AI capabilities.
Capsule Networks
A type of artificial neural network that aims to improve the efficiency and accuracy of learning by understanding spatial hierarchies between features.
Catastrophic Forgetting
The tendency of a neural network to forget previously learned information upon learning new information.
Classification
A supervised learning technique where the output is a category.
Clustering
An unsupervised learning technique where patterns discovered in the data are used to group data points.
Cortana
Cortana is a virtual assistant developed by Microsoft, available across various Microsoft products and platforms.
Cryptocurrency
Cryptocurrency is a digital or virtual currency that uses cryptography for security and operates independently of a central bank.
Curriculum Learning
A methodology in machine learning where the models are trained in a way that gradually increases the difficulty of tasks.
D
DL
Deep Learning (DL) is a subset of machine learning that uses neural networks with many layers.
DNN
Deep Neural Network (DNN) is a neural network with multiple layers between the input and output layers.
Data Augmentation
A strategy used to increase the diversity of data available for training models without actually collecting new data.
Data Mining
The process of discovering patterns and knowledge from large amounts of data.
Data Science
An interdisciplinary field that uses scientific methods, processes, algorithms, and systems to extract knowledge and insights from structured and unstructured data.
Decision Trees
A decision support tool that uses a tree-like graph or model of decisions and their possible consequences.
Deep Learning
A subset of machine learning in artificial intelligence that has networks capable of learning unsupervised from data that is unstructured or unlabeled.
Deepfake
Deepfake is a technique for creating synthetic media in which a person in an existing image or video is replaced with someone else's likeness.
Dimensionality Reduction
The process of reducing the number of random variables under consideration, by obtaining a set of principal variables.
Distributed Computing
A field of computer science that studies distributed systems.
E
Edge Computing
Edge Computing is a distributed computing paradigm that brings computation and data storage closer to the location where it is needed.
Embedding
A representation of data where elements with a similar meaning have a similar representation in vector space.
Embedding Layer
A layer in neural networks used to convert sparse categorical data into a dense embedded representation suitable for machine learning.
Ensemble Learning
A technique that combines the predictions from multiple machine learning algorithms together to make more accurate predictions than any individual model.
Ethical AI
Ethical AI refers to the development and use
Evolutionary Algorithms
Algorithms that use mechanisms inspired by biological evolution, such as reproduction, mutation, recombination, and selection.
Explainability
Explainability refers to the ability of an AI system to provide explanations for its decisions and actions in a way that is understandable to humans.
Explainable AI (XAI)
Artificial intelligence and machine learning techniques that provide human-understandable explanations of their output.
Exponential Smoothing
A rule of thumb technique for smoothing time series data, particularly for recursively applying as many as three smoothing constants.
F
Feature
An individual measurable property or characteristic of a phenomenon being observed.
Feature Engineering
The process of using domain knowledge to extract features from raw data.
Feature Extraction
The process of transforming raw data into a set of features that can be used for training a model.
Feature Scaling
A method used to normalize the range of independent variables or features of data.
Federated Learning
A machine learning setting where the model is trained across multiple decentralized devices or servers holding local data samples, without exchanging them.
Fine-tuning
A process of adjusting the parameters of an existing model to make it perform better on a specific task.
G
GAN
Generative Adversarial Network, a class of machine learning frameworks designed by two neural networks, one generating candidates and the other evaluating them.
GPT
Generative Pre-trained Transformer (GPT) is a type of deep learning model known for its ability to generate human-like text.
General Adversarial Networks (GANs)
A class of machine learning frameworks designed by unifying two neural networks, competing against each other to generate new data with the same statistics as the training set.
Genetic Algorithms
Search heuristics that mimic the process of natural selection to generate useful solutions to optimization and search problems.
Gradient Checking
A procedure used to verify the correctness of the backpropagation process by comparing the gradient computed using backpropagation against a numerically estimated gradient.
Gradient Descent
An optimization algorithm used to minimize some function by iteratively moving in the direction of steepest descent as defined by the negative of the gradient.
H
Heuristics
Techniques 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.
Hierarchical Clustering
A method of cluster analysis which seeks to build a hierarchy of clusters.
Hyperparameter
A parameter whose value is set before the learning process begins.
Hyperparameter Tuning
The process of optimizing the parameters that govern the training process of a machine learning model.
I
Imbalanced Learning
Dealing with data sets where one class is significantly underrepresented compared to others, making model training challenging.
Imputation
The process of replacing missing data with substituted values.
Inference
The process of using a trained model to make predictions.
Instance-based Learning
A family of learning algorithms that, instead of performing explicit generalization, compares new problem instances with instances seen in training.
IoMT
Internet of Medical Things (IoMT) refers to medical devices and applications connected to healthcare IT systems through the internet.
IoP
Internet of People (IoP) refers to the concept of connecting individuals to the internet through wearable devices and personal technology.
IoT
Internet of Things (IoT) refers to the network of physical devices embedded with sensors, software, and other technologies to connect and exchange data.
J
Jaccard Index
A statistic used for gauging the similarity and diversity of sample sets.
Jacobian Matrix
A matrix of all first-order partial derivatives of a vector-valued function.
Joint Learning
A learning paradigm where multiple tasks are learned at the same time, sharing representations.
K
K-means
A popular clustering algorithm that divides a set of data points into k groups based on feature similarity.
Kernel Methods
A class of algorithms for pattern analysis, whose best known member is the support vector machine.
Knowledge Distillation
A technique where a small model is trained to mimic a larger, pre-trained model.
Knowledge Graph
A knowledge base used by Google and its services to enhance its search engine's results with information gathered from a variety of sources.
L
LSTM
Long Short-Term Memory (LSTM) is a type of recurrent neural network capable of learning long-term dependencies.
Labeled Data
Data that has been tagged with one or more labels identifying certain properties or categories.
Latent Variable
Variables that are not directly observed but are rather inferred from other variables that are observed and directly measured.
Learning Rate
A hyperparameter that controls how much to change the model in response to the estimated error each time the model weights are updated.
Long Short-Term Memory (LSTM)
A special kind of RNN, capable of learning long-term dependencies.
Loss Function
A method of evaluating how well specific algorithm models the given data.
M
ML
Machine Learning (ML) is a subset of AI that focuses on getting machines to learn from data.
MLaaS
Machine Learning as a Service (MLaaS) refers to cloud-based platforms that provide machine learning tools and infrastructure.
Machine Perception
The capability of a computer system to interpret data in a manner that is similar to the way humans use their senses to relate to the world around them.
Markov Decision Processes (MDP)
A mathematical framework for modeling decision making in situations where outcomes are partly random and partly under the control of a decision maker.
Model
In machine learning, a model is a representation of what an AI system has learned from the training data.
Model Compression
Techniques to reduce the size of a machine learning model without significantly reducing its accuracy.
Multilayer Perceptron (MLP)
A class of feedforward artificial neural network (ANN).
N
NLP
Natural Language Processing (NLP) involves the interaction between computers and humans using natural language.
NN
Neural Network (NN) is a computational model inspired by the structure and function of the human brain.
Natural Language Processing (NLP)
A field of artificial intelligence that gives machines the ability to read, understand and derive meaning from human languages.
Natural Language Understanding (NLU)
A sub-discipline of natural language processing that focuses on machine reading comprehension.
Neuron
A basic unit of computation in a neural network.
Noise Reduction
The process of removing noise from a signal in data preprocessing.
Normalization
A process that changes the range of pixel intensity values.
O
Objective Function
A function that a model aims to minimize or maximize during the training process.
Optical Character Recognition (OCR)
The mechanical or electronic conversion of images of typed, handwritten or printed text into machine-encoded text.
Outlier Detection
Identifying data points that differ significantly from the majority of the data.
Overfitting
A modeling error that occurs when a function is too closely fit to a limited set of data points.
P
Partial Dependence Plot
A graphical representation of the marginal effect of a variable on a response in a statistical model.
Pooling Layer
A layer in a neural network that reduces the dimensionality of data by combining the outputs of neuron clusters.
Precision
A measure of a classifier's exactness. The higher the precision, the more accurate the classifier.
Principal Component Analysis (PCA)
A statistical procedure that uses an orthogonal transformation to convert a set of observations of possibly correlated variables into a set of values of linearly uncorrelated variables.
PyTorch
PyTorch is an open-source machine learning library developed by Facebook's AI Research lab.
Q
Quantile Regression
A type of regression analysis used in statistics and econometrics that allows the user to estimate either the conditional median or quantiles of the response variable.
Quantization
The process of reducing the number of bits that represent a number, used in model compression.
Quantum Computing
Quantum Computing is a type of computing that takes advantage of the quantum state of subatomic particles to perform operations on data.
Quantum Machine Learning
An emerging interdisciplinary research area at the intersection of quantum physics and machine learning.
R
RL
Reinforcement Learning (RL) is a type of machine learning where an agent learns to make decisions by trial and error.
RNN
Recurrent Neural Network (RNN) is a type of neural network designed to recognize patterns in sequences of data.
RPA
Robotic Process Automation (RPA) refers to the use of software robots to automate repetitive tasks and processes.
Random Forest
An ensemble learning method for classification, regression and other tasks that operates by constructing a multitude of decision trees at training time.
Recurrent Neural Network (RNN)
A class of artificial neural networks where connections between units form a directed graph along a sequence, allowing it to exhibit temporal dynamic behavior.
Regularization
The process of adding a penalty on the different parameters of the model to reduce the freedom of the model thereby overfitting.
Reinforcement Learning
A type of machine learning technique that enables an algorithm to learn through trial and error using feedback from its own actions and experiences.
Reinforcement Learning Environment
The context or setting in which a reinforcement learning model interacts and learns.
S
SVM
Support Vector Machine (SVM) is a supervised machine learning algorithm used for classification and regression analysis.
Semi-supervised Learning
A class of machine learning tasks and techniques that also make use of unlabeled data for training – typically a small amount of labeled data with a large amount of unlabeled data.
Siri
Siri is a virtual assistant developed by Apple that uses natural language processing and voice recognition technology.
Smart Contract
A Smart Contract is a self-executing contract with the terms of the agreement between buyer and seller directly written into code.
Sparse Representation
A way of representing data in a matrix form where most of the elements are zero.
Stochastic Gradient Descent (SGD)
A stochastic approximation of the gradient descent optimization and a standard algorithm for training deep learning models.
Supervised Learning
A type of machine learning algorithm that uses a known dataset (known as the training dataset) to make predictions.
T
TF
TensorFlow (TF) is an open-source machine learning framework developed by Google.
Temporal Difference Learning
A prediction method that learns by bootstrapping from the current estimate of the value function.
Tokenization
The process of converting a sequence of characters into a sequence of tokens.
Transfer Learning
Improving learning in a new task through the transfer of knowledge from a related task that has already been learned.
U
Underfitting
A model that can neither model the training data nor generalize to new data.
Unstructured Data
Information that either does not have a pre-defined data model or is not organized in a pre-defined manner.
Unsupervised Learning
A type of algorithm that learns patterns from untagged data.
Unsupervised Pre-training
Training a model on a large unsupervised dataset to learn general features that can be useful for downstream tasks.
V
VA
Virtual Assistant (VA) is an AI-driven software program that can perform tasks or services for an individual, typically through voice commands.
VR
Virtual Reality (VR) is a simulated experience that can be similar to or completely different from the real world.
Validation Set
A set of data used to assess the strength and utility of a predictive relationship.
Variance
A measure of the variability of model prediction for a given data point.
Variance Reduction
Techniques used to reduce the variance of a model's prediction error.
Variational Autoencoder (VAE)
A type of autoencoder that provides a probabilistic manner for describing observations in latent space.
Variational Inference
A technique in Bayesian machine learning that uses optimization to approximate the posterior distribution of a probabilistic model.
W
Watson
Watson is an AI system developed by IBM, known for its ability to answer questions posed in natural language.
Weak AI
Artificial intelligence that is designed and trained for a particular task.
Weight Initialization
Choosing a method to set the initial weights of an artificial neural network.
Word Embedding
A class of approaches for representing words and documents using a dense vector representation.
X
XAI
Explainable Artificial Intelligence (XAI) refers to AI systems whose decisions and actions can be understood by humans.
XAI Techniques
Methods used in explainable artificial intelligence to make the results of AI and machine learning models more understandable to humans.
XGBoost
An optimized distributed gradient boosting library designed to be highly efficient, flexible, and portable.
XOR Problem
A problem in neural network research, notable because it cannot be solved by a single layer perceptron.
Y
Yield Prediction
The application of machine learning techniques to predict the yield of crops, manufacturing processes, or other production processes.
Yielding Insights
The process of deriving actionable and meaningful information from data analysis and model predictions.
Yielding Strategy
Approaches in machine learning models that allow for dynamically adjusting actions based on changing environments or feedback loops.
Z
Z-Score Normalization
A technique used to normalize the features of a dataset by subtracting the mean and dividing by the standard deviation.
Zero-data Learning
A form of learning where the model can intelligently guess the properties of unseen categories without any associated data.
Zero-shot Learning
A machine learning technique where the model is able to recognize objects, scenes, or concepts in data that it has never explicitly been trained on.