| Internet of Things (IoT) |
A network of physical devices connected to the internet and capable of collecting, sharing, and acting on data. |
| Inference |
The process of producing predictions or judgments using a trained model. |
| Initialization |
Setting the initial parameters of a neural network before training begins. |
| Instance-based Learning |
Algorithms for comparing new problem instances with those experienced during training. |
| Inductive Learning |
A machine learning technique where the model learns general rules and patterns from specific examples or training data so the model can make accurate predictions on unseen data by applying the learned rules. |
| Input Layer |
The layer in a neural network that first receives the input data. |
| Image Recognition |
AI technology allows computers to recognize and classify objects, persons, or situations in images in the same way that humans do. |
| Image Processing |
A process using computers to enhance, analyze, or manipulate images to improve visual quality or to extract useful information from them. |
| Intelligent Agent |
A computer software that can observe its surroundings, make judgments, and take actions to achieve a certain goal. |
| Information Gain |
A metric used in decision trees to determine which characteristic (or question) is the most useful for separating data into groups. |
| Independent Component Analysis (ICA) |
A technique for isolating a complex group of signals into individual sources. |
| Imputation |
The process of replacing missing data with substituted values. |
| Isotonic Regression |
A regression technique that fits a non-decreasing function to data. |
| Invariant Representation |
Features that remain constant under various transformations, aiding in recognition tasks. |
| Integrated Gradients |
An attribution method in explainable AI that assigns importance scores to input features. |
| Imbalanced Dataset |
A dataset where some classes are significantly more frequent than others, posing challenges for model training. |
| Interpolation |
Estimating unknown values that fall within the range of known data points. |
| Intent Recognition |
Identifying the purpose behind a user's input, commonly used in natural language processing. |
| Interactive Learning |
Learning methods that involve interaction between the model and the environment or user. |
| Instance Segmentation |
A computer vision task that identifies each object instance in an image at the pixel level. |
| Interpretability |
The degree to which a human can understand the cause of a decision made by a model. |
| Inverse Reinforcement Learning |
Learning the reward function given observations of optimal behavior. |
| Imitation Learning |
Training models to perform tasks by mimicking expert behavior. |
| Intrusion Detection |
Identifying unauthorized actions in a network or system. |
| Inception Network |
A type of deep neural network architecture that uses parallel convolutional layers with different filter sizes. |
| Information Retrieval |
The process of obtaining relevant information from a large repository. |
| Independent Variable |
In modeling, a variable manipulated to observe its effect on a dependent variable. |
| Incremental Learning |
An approach where the model learns continuously from new data without forgetting previous knowledge. |
| Isolation Forest |
An algorithm for anomaly detection that isolates anomalies instead of profiling normal data. |
| Instance Weighting |
Assigning different weights to data instances, often used to handle imbalanced datasets. |
