Artificial Intelligence Concepts, Agents, and Problem Formulation

Comprehensive vocabulary flashcards covering AI introduction, history, applications, Turing Test components, AI agents and environments, PEAS framework, and problem formulation components.

Artificial Intelligence (AI)

A branch of computer science focused on creating machines and systems capable of performing tasks that normally require human intelligence, such as learning, reasoning, decision-making, and problem-solving.

Dartmouth Conference (1956)

The historical conference where the term "Artificial Intelligence" was first used, establishing AI as an official research field based on the belief that machines could learn to think like humans.

AI Winter

A period during the 1970s–1980s when AI progress slowed down, funding decreased, and public interest declined due to limited computing power and unmet expectations.

Deep Blue

The chess-playing computer system that defeated world chess champion Garry Kasparov in 1997.

Turing Test

A test proposed by Alan Turing in 1950 to evaluate machine intelligence by having a human interrogator converse via text with a human and a computer to see if the machine can successfully imitate human thinking.

Interrogator (Player C)

The human judge in the Turing Test who asks text-based questions to a computer (Player A) and a human responder (Player B) without knowing which is which, trying to determine the machine's identity.

AI Agents

Autonomous systems designed to perceive their environment, make decisions, and execute actions independently to achieve specific defined goals.

Perception (AI Agent Function)

The step in an AI agent's operation where it gathers information from its environment using sensors, user inputs, or external databases.

Decision-Making (AI Agent Function)

The stage in which an AI agent analyzes collected environmental data using rule-based logic, machine learning, or advanced techniques to determine the best action to take.

Action Execution (AI Agent Function)

The execution step where an AI agent carries out tasks based on its decisions, such as answering queries, adjusting devices, or automating workflows.

Simple Reflex Agents

The most basic type of AI agent that selects actions based solely on the current percept using condition-action (if-then) rules, completely ignoring past history and context.

Model-Based Reflex Agents

AI agents that maintain an internal model of the world to track unobservable aspects over time, enabling them to handle partially observable environments effectively.

Goal-Based Agents

AI agents that combine an internal model of the world with goal representations, utilizing search and planning algorithms to determine action sequences that reach a desired goal state.

Utility-Based Agents

AI agents that select actions to maximize expected utility or user satisfaction, allowing rational decision-making in stochastic environments where outcomes are uncertain.

Learning Agents

AI systems that autonomously interact with their environment, learn from experiences and feedback, and adapt their behaviors to improve performance over time.

Reinforcement Learning Agent

A type of learning agent that learns optimal actions by receiving rewards for favorable outcomes and penalties for unfavorable ones, working to maximize cumulative rewards.

PEAS Framework

An AI framework used to analyze and design intelligent agents by breaking down interaction into Performance Measure, Environment, Actuators, and Sensors.

Performance Measure (P in PEAS)

The objective criteria or metrics used to evaluate how successfully an AI agent is achieving its goals.

Environment (E in PEAS)

The entire surrounding context or space that an AI agent observes and interacts with while performing its tasks.

Actuators (A in PEAS)

The mechanisms or components through which an AI agent performs physical or digital actions to affect its environment.

Sensors (S in PEAS)

The components that collect environmental data and feed it to the agent so it can understand its surroundings and make informed decisions.

Fully Observable vs. Partially Observable Environment

A fully observable environment provides complete information about the current state, whereas a partially observable environment hides certain inputs or variables from direct perception.

Deterministic vs. Stochastic Environment

In a deterministic environment, action outcomes are entirely predictable; in a stochastic environment, outcomes involve uncertainty and probabilities.

Static vs. Dynamic Environment

A static environment remains unchanged while the agent is deciding on an action, whereas a dynamic environment continuously updates in real-time.

Episodic vs. Sequential Environment

An episodic environment treats each decision episode independently using only current percepts, whereas a sequential environment relies on stored percept history to dictate current decisions.

Problem Formulation

The initial step in AI problem solving that defines the problem in a structured form (Initial State, Goal State, Actions, State Space) so an AI system can search for a solution efficiently.

Components of AI Problem Formulation

The four foundational elements required to model an AI problem: Initial State (starting point), Goal State (destination), Actions (operators), and State Space (all reachable states).

Initial State

The starting condition or configuration of the environment from which an AI agent begins searching for a solution.

Goal State

The target state or condition that an AI agent wants to reach, marking the successful completion of the search process.

Actions (Operators)

The set of valid operations available to an AI agent that transition the environment from one state to another.

State Space

The total collection of all possible states that can be reached from the initial state by applying valid sequences of actions.

Transition Model

A function in AI problem formulation that describes what state results from taking a specific action in a given state.

Goal Test

A test procedure that determines whether a given state matches the desired goal state.

Path Cost

A numeric function that assigns a numeric cost (such as distance, time, fuel, or energy) to a path through the state space.

AI Application Problem Formulations (Group 1)

Formulation mappings for real-world applications including Google Maps, Vacuum Cleaner, Chess, Self-Driving Car, and Medical Diagnosis across initial states, goal states, actions, and state spaces.

AI Application Problem Formulations (Group 2)

Formulation specifications for Warehouse Robot, Sudoku Solver, Spam Detection, Face Recognition, and Delivery Drone.