Fundamentals of AI and Applications Practice Flashcards

Vocabulary flashcards covering Modules 1 to 5 of the Fundamentals of AI and Applications course, including agent types, search algorithms, machine learning models, robotics, and ethical considerations.

Artificial Intelligence (AI)

A subfield of computer science concerned with building smart systems capable of performing tasks that typically require human intelligence, involving algorithms and models that allow machines to learn, reason, and adapt.

Narrow AI (Weak AI)

AI systems designed and trained to complete a specific, specialized task without generalized intellectual capabilities, operating under a limited context like chess computers or facial recognition.

Strong AI (General AI / AGI)

A theoretical form of AI where a machine possesses human-level cognitive capabilities across a broad spectrum of domains, allowing it to reason, learn, and apply intelligence flexibly to unfamiliar tasks.

Reactive Machines

The most fundamental type of AI system that possesses no memory or ability to store past experience, reacting immediately to current situations based on pre-programmed rules.

Goal-Based Agent

An agent that selects its actions based on achieving desirable goals, combining environmental knowledge with explicit instructions to evaluate which actions bring it closer to its target.

Fully Observable Environment

An environment where the agent's sensors can access the complete state of the environment at any given point, such as in the game of Chess.

Deterministic Environment

An environment where the next state is completely determined by the current state and the action executed by the agent, such as in Sudoku.

Episodic Environment

An environment where the agent's experience is divided into independent atomic episodes, and actions taken in one episode do not affect subsequent ones.

Static Environment

An environment that does not change while the agent is deliberating, such as a crossword puzzle.

Discrete Environment

An environment featuring a finite, distinct number of actions and possible states, typical of board games.

Uniform Cost Search (UCS)

An uninformed search algorithm that expands nodes based on their path cost from the root using a priority queue ordered by the cumulative path cost $g(n)$.

Super AI (ASI)

A theoretical level of AI that surpasses human capability across every cognitive measure.

Limited Memory AI

AI systems that can retain historical data for a brief duration to guide choices, such as autonomous cars tracking nearby vehicles.

Theory of Mind

A theoretical category of AI capable of understanding social and psychological factors along with human emotions.

Self-Aware AI

A theoretical category of AI that possesses full self-consciousness and an independent internal identity.

Deductive Reasoning

A top-down approach that moves from general premises to a specific certain conclusion; if premises are true, the conclusion must be true.

Inductive Reasoning

A bottom-up approach that moves from specific observations to general probabilistic theories where the conclusion could be proven false.

Simple Reflex Agent

An agent that selects actions exclusively based on the current state via direct Condition-Action rules, ignoring historical context.

Bidirectional Search

A graph traversal algorithm that runs two simultaneous searches forward from the start and backward from the goal, reducing time complexity to $O(b^{d/2})$.

Greedy Best-First Search

An informed search algorithm that expands nodes solely based on the lowest heuristic score $f(n) = h(n)$.

A* Search

An informed search algorithm combining historical path cost and heuristic estimated cost to the goal using the formula $f(n) = g(n) + h(n)$.

Breadth-First Search (BFS)

An uninformed traversal strategy that explores all vertices at a given depth level using a First-In-First-Out (FIFO) queue before moving deeper.

Classification

A supervised learning task where the objective is to predict a discrete qualitative category label from a predefined finite set.

Regression

A supervised learning task where the objective is to predict a continuous quantitative numerical value across an infinite range.

Structured Data

Highly organized data following a rigid schema that fits into relational databases, such as SQL tables and CSV logs.

Unstructured Data

Raw data lacking predefined internal organization, such as video feeds, audio recordings, and social media posts.

Nominal Data

Categorical data that has no inherent ranking or logical order, acting purely as labels like eye color or car brands.

Ordinal Data

Categorical data with a clear hierarchical ranking or natural order, such as customer satisfaction levels or education levels.

Supervised Learning

A type of machine learning that uses labeled data consisting of pairs of inputs and corresponding true output targets to train a model.

Unsupervised Learning

A type of machine learning that processes unlabeled data to discover hidden structural patterns, associations, or groupings.

Five V's of Big Data

The dimensions defining Big Data: Volume, Velocity, Variety, Veracity, and Value.

Linear Regression

A model of the linear relationship between a continuous dependent target variable and independent variables using the equation $Y = \beta_0 + \beta_1 X + \nu$.

Logistic Regression

A classification algorithm that predicts the probability of a binary outcome using the Sigmoid function: $P(Y=1|X) = \frac{1}{1 + e^{-(\beta_0 + \beta_1 X)}}$.

K-Means

An iterative unsupervised clustering algorithm that groups a dataset into $K$ distinct clusters based on feature similarity and Euclidean distance.

Support Vector Machine (SVM)

A model that finds an optimal separating hyperplane in an n-dimensional space that maximizes the margin between different data classes.

Decision Tree

A flowchart-like tree structure where internal nodes represent feature tests and leaf nodes represent final class labels.

Random Forest

An ensemble learning technique that builds a collection of individual decision trees and uses a majority vote for the final classification.

Density-Based Clustering (DBSCAN)

A technique that defines clusters as continuous high-density regions separated by low-density areas, capable of discovering arbitrary shapes.

Naive Bayes

A probabilistic classification algorithm based on Bayes' Theorem assuming all features are independent: $P(C|X) = \frac{P(X|C) \times P(C)}{P(X)}$.

Industrial Robots

Stationary articulating arms used in manufacturing for tasks like spot welding, precision assembly, and heavy lifting.

Actuators

The 'muscles' of a robot that convert electrical energy into physical movement, such as DC motors and hydraulic pumps.

SLAM

Simultaneous Localization and Mapping; algorithms that allow a robot to map an unfamiliar environment while tracking its own location within it.

AI as a Service (AaaS)

The offering of AI capabilities via cloud-based providers, making advanced models accessible through APIs without heavy upfront infrastructure costs.

Generative Adversarial Networks (GAN)

An architecture where two neural networks, a Generator and a Discriminator, compete in a feedback loop to create realistic synthetic data.

Forward Chaining

A data-driven strategy in inference engines that starts with known facts and applies IF-THEN rules to infer new facts toward a conclusion.

Backward Chaining

A goal-driven reasoning strategy that starts with a specific hypothesis and works backward to see what conditions support it.

Expert System

An AI application mimicking a human expert's decision-making ability using a knowledge base and an inference engine.

Ethical AI Pillars

The core values ensuring AI safety and fairness: Fairness, Accountability, Transparency (XAI), and Privacy.