Artificial Intelligence Fundamentals

A complete set of vocabulary flashcards covering AI concepts, machine learning techniques, neural networks, computer vision, natural language processing, and responsible AI principles based on the lecture notes.

Artificial Intelligence (AI)

Technology that enables computers to perform tasks that normally require human intelligence, such as understanding language, recognizing images, making predictions, finding patterns, and solving problems.

Artificial Intelligence, Machine Learning, and Deep Learning Relationship

A nested hierarchy where Artificial Intelligence is the broad field enabling machine problem-solving; Machine Learning is a subset focused on systems learning automatically from experience; and Deep Learning is a specialized machine learning technique using multi-layered artificial neural networks.

Machine Learning (ML)

A method of creating AI systems that learn patterns automatically from data and experience instead of being explicitly programmed for every situation.

Model

A trained system that has learned patterns from data during training and can use those learned patterns to make predictions, decisions, or generate outputs.

Algorithm

The specific method or process used to learn from data, which combines with training data to produce a trained model.

Deep Learning (DL)

A specialized form of machine learning that uses multi-layered artificial neural networks (neurons) to learn complex patterns, mimicking the human brain.

Generative AI

A type of AI that creates new content—such as text, images, audio, video, and code—based on user prompts or inputs.

Training

The phase in which an AI model analyzes large amounts of data and adjusts its internal parameters to learn underlying patterns.

Inference

The phase in which a trained model applies its learned patterns to new, unseen data or prompts to generate predictions or content.

Supervised Learning

A machine learning category where a model learns from labeled data that includes correct answers, finding patterns between inputs and target answers.

Regression

A supervised learning task that predicts a continuous numerical value, such as measuring rainfall in inches or millimeters.

Classification

A supervised learning task that predicts a discrete category or class label, such as identifying whether a photograph contains a cat.

Unsupervised Learning

A machine learning category where a model discovers patterns, relationships, or natural groupings in unlabeled data without target answers.

Clustering

An unsupervised learning technique that groups similar data points together without pre-defined category labels.

Reinforcement Learning

A machine learning category where an agent learns through trial and error by taking actions within an environment to maximize cumulative rewards.

Large Language Models (LLMs)

AI systems built on neural networks with millions or billions of parameters, trained on vast text datasets to understand and generate human language.

Neural Networks

Computer models inspired by information processing in the human brain, composed of connected layers of nodes (neurons) designed to recognize data patterns.

Parameters

Internal values learned by a model from training data that dictate how inputs are processed to calculate predictions.

Hyperparameters

Settings chosen prior to training that govern how the model learns, including learning rate, epochs, batch size, and network layers.

Learning Rate

A hyperparameter that controls how quickly or aggressively a model adjusts its internal parameters during training iterations.

Epochs

A hyperparameter specifying the total number of complete passes the learning algorithm makes through the entire training dataset.

Batch Size

A hyperparameter defining the quantity of data samples processed at one time before updating the model parameters.

Natural Language Processing (NLP)

A field of AI focused on enabling computers to understand, interpret, analyze, and generate natural human language.

Computer Vision

An area of AI dealing with the visual analysis of visual inputs, such as photographs, videos, and live camera feeds.

Image Classification

A computer vision task where a model is trained on labeled images to predict the primary subject label of unlabeled images.

Object Detection

A computer vision task where a model is trained to identify and locate specific objects within an image.

Semantic Segmentation

An advanced computer vision technique that classifies individual pixels in an image to precisely define object boundaries rather than using bounding boxes.

Multi-modal Models

AI models that combine multiple types of data inputs, such as combining visual features with text descriptions to produce comprehensive descriptions.

Agents

Software applications built on generative AI that reason over natural language, use external tools to automate tasks, and respond to contextual conditions.

Optical Character Recognition (OCR)

A foundational document analysis technology that detects and identifies the precise location of text in an image.

Speech Recognition

The ability of an AI system to hear, process, and accurately interpret spoken human language.

Speech Synthesis

The ability of an AI system to generate and vocalize text or words as spoken audio language.

Responsible AI

A framework and practice for developing AI systems with built-in guardrails to minimize risks of generating harmful, illegal, or offensive content.

Fairness

A principle of Responsible AI requiring developers to minimize bias in training data and test systems to ensure unbiased treatment across demographic groups.

Reliability and Safety

A principle of Responsible AI emphasizing risk mitigation to account for the probabilistic, non-infallible nature of AI applications.

Privacy and Security

A principle of Responsible AI ensuring training data containing personal details is kept secure and trained models cannot expose private information.

Inclusiveness

A principle of Responsible AI striving to make AI solutions and their life-improving benefits accessible to all users without exclusion.

Transparency

A principle of Responsible AI focusing on making users explicitly aware of how an AI system functions, how it operates, and its limitations.

Accountability

A principle of Responsible AI holding developers and distributing organizations answerable for the real-world impact and governance of their AI deployments.