Glossary - AI Terminology

Forward Propagation

The process of passing input data through a neural network to generate an output.

Types of Forward Propagation

Fully Connected Forward Propagation - Data moves through dense layers.
Convolutional Forward Propagation - Used in CNNs to extract spatial features.

Example

Used in deep learning models to compute predictions.

Forward Selection

A feature selection method that iteratively adds the most significant features to improve model performance.

Types of Forward Selection

Greedy Forward Selection - Adds features based on immediate improvement.
Stepwise Forward Selection - Allows removing previously selected features.

Example

Used in regression models to optimize feature selection.

Fractional Differencing

A technique used in time series analysis to achieve stationarity while preserving long-term dependencies.

Types of Differencing

Integer Differencing - Uses full lag differences.
Fractional Differencing - Uses partial differencing for smoother results.

Example

Used in financial market analysis for volatility modeling.

Frame Problem in AI

A challenge in AI reasoning where specifying all unchanged aspects of a system becomes computationally expensive.

Types of Frame Problem Solutions

Circumscription - Assumes only necessary changes occur.
Event Calculus - Models events over time explicitly.

Example

Used in automated planning and robotics.

Free Energy Principle

A theory in AI and neuroscience stating that intelligent systems minimize uncertainty in their predictions.

Types of Free Energy Models

Variational Free Energy - Used in probabilistic inference.
Predictive Coding - Reduces prediction errors iteratively.

Example

Used in Bayesian machine learning models.

Frequency Filtering

A technique that manipulates frequency components of signals for feature extraction.

Types of Frequency Filtering

Low-Pass Filtering - Removes high-frequency noise.
High-Pass Filtering - Enhances edges in images.

Example

Used in image processing and speech recognition.

Frequent Itemset Mining

A data mining technique used to find patterns and associations in large datasets.

Types of Frequent Itemset Mining

Apriori Algorithm - Generates frequent itemsets using support thresholds.
Eclat Algorithm - Uses depth-first search for efficiency.

Example

Used in retail market analysis for recommendation systems.

Frugal Learning

A machine learning approach that focuses on achieving high performance with minimal computational resources.

Types of Frugal Learning

Few-Shot Learning - Learns from very limited data.
Incremental Learning - Updates knowledge over time.

Example

Used in IoT applications with low-power devices.

Functional Neural Networks

A type of neural network where weights and activations are represented as continuous functions.

Types of Functional Neural Networks

Fourier Neural Networks - Uses Fourier transformations.
Neural ODEs - Defines neural networks as differential equations.

Example

Used in physics-informed machine learning for solving differential equations.

Fused Lasso

A variation of Lasso regression that encourages sparsity and smoothness in feature selection.

Types of Lasso Methods

Standard Lasso - Performs feature selection via shrinkage.
Fused Lasso - Adds smoothness constraints between features.

Example

Used in genomics for identifying key biomarkers.

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.