Image Augmentation
A technique used to artificially expand training datasets by applying transformations to images.
Types of Image Augmentation
- Geometric Transformations - Includes rotation, scaling, and flipping.
- Color Transformations - Adjust brightness, contrast, and saturation.
Example
Used in deep learning models for image classification.
Image Captioning
The process of generating textual descriptions for images using machine learning models.
Types of Image Captioning Models
- Encoder-Decoder Models - Use CNNs and RNNs.
- Transformer-based Models - Use self-attention mechanisms.
Example
Used in accessibility applications for visually impaired users.
Image Classification
The task of assigning labels to images based on their content.
Types of Image Classification
- Single-Label Classification - Assigns one label per image.
- Multi-Label Classification - Allows multiple labels per image.
Example
Used in facial recognition systems.
Image Recognition
A technology that identifies objects, people, and scenes in images using AI.
Types of Image Recognition
- Object Detection - Detects and classifies objects.
- Facial Recognition - Identifies human faces.
Example
Used in security systems for biometric authentication.
Image Segmentation
A technique that partitions images into multiple segments for better analysis.
Types of Image Segmentation
- Semantic Segmentation - Classifies each pixel into a category.
- Instance Segmentation - Differentiates individual objects of the same category.
Example
Used in medical imaging to detect tumors.
Imbalanced Data Handling
A set of techniques to manage datasets where one class is significantly larger than another.
Types of Techniques
- Oversampling - Duplicating minority class examples.
- Undersampling - Reducing majority class examples.
Example
Used in fraud detection where fraudulent cases are rare.
Incremental Learning
A machine learning approach where models learn new data without forgetting previous knowledge.
Types of Incremental Learning
- Online Learning - Continuously updates the model with new data.
- Transfer Learning - Applies previously learned knowledge to new tasks.
Example
Used in adaptive spam filters.
Independent Component Analysis (ICA)
A statistical technique used to separate mixed signals into independent components.
Types of ICA Methods
- FastICA - An efficient iterative algorithm.
- Infomax ICA - Maximizes mutual information.
Example
Used in biomedical signal processing, such as EEG analysis.
Inductive Learning
A learning paradigm where models generalize from specific training examples to unseen data.
Types of Inductive Learning
- Supervised Learning - Learns from labeled data.
- Unsupervised Learning - Finds patterns in unlabeled data.
Example
Used in spam detection models.
Types of Information Bottleneck Methods
- Variational Information Bottleneck - Uses probabilistic models.
- Deterministic Bottleneck - Uses neural network constraints.
Example
Used in deep learning for efficient representation learning.
Types of Information Gain Calculations
- Entropy-Based - Uses Shannon entropy.
- Gini Index - Measures impurity instead of entropy.
Example
Used in decision tree algorithms like ID3 and C4.5.
Types of Information Retrieval
- Text-Based Retrieval - Searches text documents.
- Multimedia Retrieval - Searches images, audio, and video.
Example
Used in search engines like Google.
Instance-Based Learning
A type of machine learning that memorizes and compares new data to previously seen examples.
Types of Instance-Based Learning
- k-Nearest Neighbors (k-NN) - Classifies based on closest neighbors.
- Case-Based Reasoning - Uses past cases to solve new problems.
Example
Used in recommendation systems.
Integrated Gradients
An explainability technique for deep learning models that attributes predictions to input features.
Types of Attribution Methods
- Saliency Maps - Highlights important regions in input.
- SHAP Values - Provides feature importance scores.
Example
Used in interpreting image classification models.
Interpolation in Machine Learning
The process of estimating unknown values within the range of known data points.
Types of Interpolation
- Linear Interpolation - Connects two points with a straight line.
- Polynomial Interpolation - Uses higher-degree polynomials.
Example
Used in missing data imputation.
Intrinsic Dimensionality
The minimum number of variables needed to describe a dataset accurately.
Types of Dimensionality Reduction Techniques
- Principal Component Analysis (PCA) - Reduces dimensions using eigenvectors.
- t-SNE - Preserves local structure in high-dimensional data.
Example
Used in feature selection for deep learning.
Invariant Representations
Feature representations that remain unchanged under certain transformations.
Types of Invariance
- Translation Invariance - Feature remains the same when shifted.
- Rotation Invariance - Feature remains unchanged under rotation.
Example
Used in convolutional neural networks (CNNs).
Iterative Deepening Search (IDS)
A search algorithm that combines depth-first and breadth-first strategies for optimal searching.
Types of Search Algorithms
- Depth-First Search - Explores deep paths first.
- Breadth-First Search - Explores all nodes at a given level.
Example
Used in game-playing AI like chess engines.
Isotropic Gaussian Distribution
A Gaussian distribution where all dimensions have equal variance.
Types of Gaussian Distributions
- Isotropic Gaussian - Same variance in all directions.
- Anisotropic Gaussian - Different variances along different axes.
Example
Used in generative models like Variational Autoencoders (VAEs).
Iterative Refinement
A process in machine learning where models iteratively improve by refining their predictions.
Types of Iterative Refinement Methods
- Expectation-Maximization (EM) - Used in probabilistic models.
- Gradient Descent - Iteratively improves model parameters.
Example
Used in training deep learning models.
Importance Sampling
A technique used to estimate properties of a distribution using a different probability distribution.
Types of Importance Sampling
- Self-Normalized Importance Sampling - Adjusts weights dynamically.
- Adaptive Importance Sampling - Modifies sampling strategy over time.
Example
Used in Monte Carlo methods for reinforcement learning.
Implicit Models
Machine learning models that define probability distributions without explicitly modeling them.
Types of Implicit Models
- Generative Adversarial Networks (GANs) - Learn distributions implicitly.
- Energy-Based Models - Define distributions using energy functions.
Example
Used in image generation and style transfer.
Imputation Methods
Techniques for handling missing data in datasets.
Types of Imputation
- Mean/Median Imputation - Replaces missing values with mean or median.
- K-Nearest Neighbors Imputation - Predicts missing values based on similar data points.
Example
Used in handling missing patient records in medical datasets.
Independent and Identically Distributed (IID) Assumption
An assumption that data points are drawn from the same probability distribution and are independent of each other.
Types of Data Distributions
- IID Data - Follows the IID assumption.
- Non-IID Data - Shows dependency between samples.
Example
Used in statistical modeling and machine learning training.
Independent Variables
Variables in a dataset that are manipulated to observe their effect on dependent variables.
Types of Independent Variables
- Continuous Independent Variables - Take numerical values.
- Categorical Independent Variables - Represent discrete categories.
Example
Used in regression models to predict house prices.
Inductive Bias
Assumptions made by a machine learning model to generalize beyond the training data.
Types of Inductive Bias
- Preference Bias - Prefers simpler hypotheses.
- Restriction Bias - Limits the set of hypotheses.
Example
Used in neural networks to learn hierarchical features.
Inference in Machine Learning
The process of using a trained model to make predictions on new data.
Types of Inference
- Batch Inference - Processes multiple inputs at once.
- Real-Time Inference - Predicts data as it arrives.
Example
Used in speech recognition systems like Siri.
Types of Information Measures
- Shannon Entropy - Measures uncertainty.
- Kullback-Leibler Divergence - Measures difference between two distributions.
Example
Used in feature selection and clustering.
Initialization Methods in Neural Networks
Techniques to initialize weights in neural networks to improve training efficiency.
Types of Initialization
- Xavier Initialization - Distributes weights evenly.
- He Initialization - Optimized for ReLU activations.
Example
Used in training deep learning models efficiently.
Instance Segmentation
A computer vision task that identifies and differentiates each object instance in an image.
Types of Segmentation
- Semantic Segmentation - Labels pixels by category.
- Instance Segmentation - Differentiates individual objects.
Example
Used in autonomous vehicle perception systems.
Integrated Learning
A learning approach that combines multiple models or data sources to improve predictive accuracy.
Types of Integrated Learning
- Ensemble Learning - Combines multiple models.
- Multi-Modal Learning - Merges different types of data.
Example
Used in self-driving cars by combining vision, radar, and sensor data.
Interaction Effects
Phenomenon where the effect of one feature depends on the value of another feature.
Types of Interaction Effects
- Synergistic Interaction - Features enhance each other.
- Antagonistic Interaction - Features counteract each other.
Example
Used in feature engineering for predictive modeling.
Interpretable Machine Learning
The study of making machine learning models understandable and explainable.
Types of Interpretability Methods
- Intrinsic Interpretability - Models that are inherently explainable (e.g., Decision Trees).
- Post-Hoc Interpretability - Methods applied after model training (e.g., LIME, SHAP).
Example
Used in finance and healthcare to explain AI decisions.
Types of Predictions
- Interpolation - Estimating values within the data range.
- Extrapolation - Estimating values beyond the known range.
Example
Used in time-series forecasting.
Interval-Based Learning
A method where models learn from data segments rather than individual points.
Types of Interval-Based Learning
- Fixed-Interval Learning - Uses pre-defined intervals.
- Adaptive-Interval Learning - Adjusts intervals dynamically.
Example
Used in financial modeling to analyze stock trends.
Invariant Risk Minimization
A learning framework that aims to find features that generalize across different environments.
Types of Invariant Learning Approaches
- Empirical Risk Minimization - Minimizes average error.
- Structural Risk Minimization - Balances model complexity and error.
Example
Used in domain adaptation for robust AI models.
Irregular Time Series
Time series data where observations occur at uneven time intervals.
Types of Time Series
- Regular Time Series - Data collected at fixed intervals.
- Irregular Time Series - Data collected at varying intervals.
Example
Used in medical monitoring for patient heart rates.
Iterative Bootstrapping
A resampling technique where models are repeatedly trained on different bootstrapped samples.
Types of Bootstrapping
- Parametric Bootstrapping - Assumes a known data distribution.
- Non-Parametric Bootstrapping - Does not assume a fixed distribution.
Example
Used in ensemble methods like Bagging.
Iterative Pruning
A method of progressively removing less important neurons or parameters in a neural network.
Types of Pruning
- Weight Pruning - Removes small-weight connections.
- Neuron Pruning - Eliminates entire neurons.
Example
Used in model compression for mobile AI applications.
Iterative Reinforcement Learning
A training method where reinforcement learning agents improve through iterative updates.
Types of Iterative Learning Approaches
- Policy Iteration - Alternates between policy evaluation and improvement.
- Value Iteration - Updates value functions to find the best policy.
Machine Learning (ML)
ML is a subset of AI that enables machines to learn patterns from data and make predictions or decisions without explicit programming.
Types of ML
- Supervised Learning
- Unsupervised Learning
- Reinforcement Learning
Example
Spam detection in emails using classification models.
Deep Learning (DL)
DL is a subset of ML that uses artificial neural networks to process complex data and perform high-level computations.
Example
Image recognition in self-driving cars.
Generative AI (Gen AI)
Gen AI refers to AI models that generate new content, including text, images, and code, using trained knowledge bases.
Example
AI models like ChatGPT and Stable Diffusion that generate text and images.