Introduction
AI tools like LLMs and Generative AI are evolving at lightning speed. To keep up, you need to understand the language. This AI & Machine Learning glossary is your go-to reference, breaking down essential terms into simple definitions for learners at any level.
Your quick-start guide to the core principles of artificial intelligence and machine learning. These terms explain the “what” and “why” behind the field.
Artificial Intelligence (AI)
Artificial Intelligence is the science of making machines perform tasks that normally need human intelligence. It includes things like understanding language, recognizing images, making decisions, or learning from experience.
Machine Learning (ML)
Machine Learning is a branch of AI where computers learn patterns from data instead of being given exact instructions. The more data they see, the better they get at making predictions or finding patterns.
Supervised Learning
Supervised Learning is when a model is trained on labeled data, meaning the input comes with the correct answer. It learns the mapping between inputs and outputs, similar to how a student practices with solved examples before taking a test.
Unsupervised Learning
Unsupervised Learning is when the model works with data that has no labels. The goal is to find hidden patterns or groupings, like sorting books by theme without knowing their categories in advance.
Reinforcement Learning
Reinforcement Learning is a way of training models by rewarding good decisions and penalizing bad ones. It is like training a pet, where the system learns through trial and error to maximize rewards over time.
Data for Machine Learning
Every model starts with data. This section breaks down key concepts about datasets, features, labels, and the steps required to clean and prepare data for training.
Dataset
A dataset is a collection of data used to train or test a machine learning model. It usually comes in rows and columns, like a table, where each row represents an example.
Features
Features are the individual input variables used by the model to make predictions. For example, in predicting house prices, features could be size, location, or number of rooms.
Labels
Labels are the correct answers or outcomes the model is trying to predict. In supervised learning, each input has a label, such as “cat” or “dog” in an image dataset.
Data Preprocessing
Data preprocessing is the step of cleaning and preparing raw data before feeding it into a model. This can include handling missing values, normalization, or removing duplicates.
Feature Engineering
Feature engineering is the process of creating new useful features or modifying existing ones to help the model learn better. It often improves accuracy more than changing the algorithm itself.
Algorithms & Models
The heart of machine learning. Here you’ll find definitions of popular algorithms and model types; the techniques that transform raw data into predictions.
Linear Regression
Linear regression is a simple algorithm that finds the best straight line to predict a continuous value, such as price or temperature, from input data.
Decision Tree
A decision tree is a model that splits data into branches based on rules, ending in predictions at the leaves. It is easy to understand and visualize.
Random Forest
Random forest combines many decision trees to make stronger predictions. Each tree sees part of the data, and the final result is decided by averaging or majority voting.
Support Vector Machine (SVM)
SVM is a model that tries to find the best boundary that separates data points of different classes. It works well for both linear and non-linear problems.
K-Means Clustering
K-means is an unsupervised algorithm that groups data into a chosen number of clusters based on similarity. It is often used for segmentation.
Model Training & Evaluation
From fitting a model to measuring its performance, this section covers how learning happens and how success is judged, including accuracy, precision, recall, and more.
Training Set
The training set is the portion of data used to teach the model by adjusting its internal parameters.
Validation Set
The validation set is used during training to check performance and tune parameters without touching the final test data.
Test Set
The test set is used only at the end to measure how well the trained model performs on unseen data.
Accuracy
Accuracy is a metric that measures how many predictions were correct out of all predictions. It works best when classes are balanced.
Cross-Validation
Cross-validation is a method of splitting data into multiple parts to ensure the model’s performance is consistent and not just luck on one split.
Neural Networks & Deep Learning
Go deeper into AI with the architectures inspired by the human brain. These terms explain layers, activation functions, CNNs, RNNs, and other advanced building blocks.
Perceptron
The perceptron is the simplest type of neural network unit that takes inputs, applies weights, and produces an output. It inspired modern deep learning.
Activation Function
An activation function decides if a neuron should fire and introduces non-linearity, allowing networks to learn complex patterns. Common examples include ReLU and Sigmoid.
Convolutional Neural Network (CNN)
CNNs are networks designed to process images by scanning them in small parts, making them powerful for vision tasks like recognition.
Recurrent Neural Network (RNN)
RNNs are networks that handle sequences by remembering past information. They are often used for text, speech, and time series.
Backpropagation
Backpropagation is the algorithm used to train neural networks by adjusting weights through error feedback. It’s like learning from mistakes step by step.
Generative AI & Transformers
The latest wave of AI. This section focuses on transformers, attention mechanisms, diffusion models, and the tools behind LLMs and creative AI applications.
Transformer
A transformer is a type of neural network architecture that processes input in parallel and uses attention to focus on important parts. It powers most modern AI models.
Attention Mechanism
Attention is a technique that lets models focus on the most relevant parts of the input. It is like highlighting important sentences in a book while reading.
Large Language Model (LLM)
LLMs are very large AI models trained on huge amounts of text data. They can understand and generate human-like language, powering tools like ChatGPT.
Diffusion Model
Diffusion models generate new data, like images, by gradually refining random noise into meaningful content. They are behind tools like Stable Diffusion.
Generative Adversarial Network (GAN)
A GAN uses two models — one generates fake data while the other tries to detect if it is real. Through competition, they improve until realistic outputs are created.
AI Ethics & Responsibility
AI is powerful, but it must also be safe and fair. This section covers bias, fairness, transparency, explainability, and the ethical challenges of modern AI.
Bias
Bias in AI happens when a model reflects unfair patterns from its training data, leading to unequal outcomes.
Fairness
Fairness means ensuring AI systems treat all groups equally without discrimination.
Transparency
Transparency is about making AI systems understandable, so users know how decisions are made.
Explainability
Explainability focuses on tools and methods that help humans understand why a model gave a certain prediction.
Responsible AI
Responsible AI is the practice of building and using AI in ways that are ethical, safe, and aligned with human values.
Conclusion
This glossary is meant to be a practical reference as you explore the world of Artificial Intelligence and Machine Learning. The field is growing fast, with new terms and concepts appearing every year. Use this guide as a starting point, and revisit it whenever you come across unfamiliar jargon. As AI continues to evolve, staying familiar with the language of the field will help you understand ideas more clearly and connect them to real-world applications.
Want to Learn More or Have Questions?
For any inquiries, contact me here.
Explore more tutorials and guides in the Blog section.
