Intelligent Agents

Flashcards covering concepts from the Intelligent Agents lecture.

Agent

Something that acts in an environment, made up of a body and a controller.

Sensors

Part of an agent that converts stimuli into percepts.

Actuators

Part of an agent, also called effectors, that convert commands into actions in the environment.

Embodied Agent

An agent that has a physical body.

Robot

An artificial purposive embodied agent.

Rational Agent

An entity that perceives its environment and acts according to some rules to achieve its goals.

Rationality

A rational agent chooses whichever action maximises the expected value of the performance measure given the percept sequence to date.

Performance Measure (Automated Taxi/Cab Example)

Safety, destination, profits, legality, comfort, etc.

Environment (Automated Taxi/Cab Example)

Streets, traffic, pedestrians, weather, etc.

Actuators (Automated Taxi/Cab Example)

Steering, accelerator, brake, horn, speaker/display, etc.

Sensors (Automated Taxi/Cab Example)

Video, accelerometers, gauges, engine sensors, keyboard, GPS, etc.

Performance Measure (Internet Shopping Agent Example)

Price, quality, appropriateness, efficiency.

Environment (Internet Shopping Agent Example)

Web sites, vendors, shippers.

Actuators (Internet Shopping Agent Example)

User interface controls, URL links, web forms

Sensors (Internet Shopping Agent Example)

HTML pages (text, graphics, scripts)

Simplest Environment

An environment that is fully observable, deterministic, episodic, static, discrete and single-agent.

Most Real Situations

An environment that is partially observable, stochastic, sequential, dynamic, continuous and multi-agent.

Perfect Rationality

The agent reasons about the best action without taking into account its limited computational resources

Bounded Rationality

The agent decides on the best action that it can find given its computational limitations.

Simple Reflex Agents

Select action on the basis of only the current percept.

Reflex Agents with State (Model-Based Reflex Agent)

Maintain internal state to tackle partially observable environments

Goal-Based Agents

The agent needs a goal to know which situations are desirable, often investigated in search and planning research. Future is taken into account.

Utility-Based Agents

Certain goals can be reached in different ways; some ways are better and have a higher utility. A utility function maps a sequence of states onto a real number

Learning Element

Introduce improvements in performance element.

Critic

Provides feedback on agent's performance based on fixed performance standard

Performance Element

Selecting actions based on percepts; Corresponds to the previous agent programs.

Problem Generator

Suggests actions that will lead to new and informative experiences.