| Adaptive Learning |
An educational system that employs artificial intelligence to personalize learning experiences by tailoring curriculum and assessments to specific students' needs and performance levels |
| Adversarial Attack |
A technique for fooling machine learning models by introducing modest, purposeful modifications to input data, which is frequently used to assess model resilience. |
| Agent |
An AI software model that can execute specific tasks autonomously on behalf of a user with or without human interaction. |
| Algorithm |
A set of rules or instructions that an AI model follows to learn from data and make judgments. |
| Algorithmic Bias |
The existence of systematic and unfair discrimination in the outcomes produced by AI systems, generally due to biased training data or model design. |
| Artificial General Intelligence (AGI) |
A theorized form of artificial intelligence capable of performing any cognitive function that a person can. |
| Artificial Intelligence |
Computer systems capable of simulating human intelligence processes. |
| Artificial Narrow Intelligence (ANI) |
AI systems that specialize in a single task or a limited set of tasks. |
| Artificial Neural Network (ANN) |
A computational model based on biological neural network structure that is employed in a variety of machine learning applications. |
| Artificial Superintelligence (ASI) |
A fictional artificial intelligence that outperforms humans in all areas, including creativity and social abilities. |
| AUC (Area Under the Curve) |
A numerical metric that determines a binary classification model's overall ability to decern between positive and negative classes. It is measured by the area below the receiver operating characteristic (ROC) curve. An AUC of 1.0 represents perfect classification, whereas an AUC of 0.5 indicates performance equivalent to random guessing. The closer the AUC is to one, the better well the model predicts the proper class. |
| Augmented Reality (AR) |
A technique that superimposes digital information, like as images or data, on the real world, to improve the user's impression of reality |
| Autoencoder |
A type of neural network used for unsupervised learning that learns efficient representations (encoding) of data, typically for dimensionality reduction or noise reduction. |
| Automated Machine Learning (AutoML) |
A system that simplifies and accelerates the process of developing, selecting, and optimizing machine learning models for real-world applications. |
| Automation |
The application of technology to accomplish activities with little or no human participation, resulting in enhanced efficiency and production. |
| Autonomous Vehicle |
A self-driving car or other vehicle that uses AI and sensor systems to navigate and operate without human input. |
