AI Midterm 1

Romania Example, Formulate Goal

Be in Bucharest

Autonomous behavior

Behavior is determined by its own experience

Sensors (Automated taxi driver)

Cameras, microphone, sonar, speedometer, gps, odometer

Perceptual sequence

Complete history of everything agent has perceived

Discrete vs continous

Distinct, clearly defined percepts and actions (Chess)

Measuring Performance: Completeness

Does it find a solution when one exists?

• Agent program

- Implements the agent function for an agent The agent function is an abstract mathematical description; the agent program is a concrete implementation, running within some physical system.

• Problem-solving agent

A kind of goal-based agent that decides what to do by finding sequences of actions that lead to desirable states

What is a solution

A sequence of operators leading from the initial state to a goal state. Optimal solution has lowest path cost

Goal

A set of states in which the goal is satisfied

Performance Measure

A way to evaluate the agent's success When to evaluate is also important(Time span)

Agent Percepts

Agent's perceptual inputs at any given instant

Agent = Architecture + Program What is architecture and program

Architecture is the computing device. Program implements agent function mapping of percepts to actions

Static vs Dynamic

Environment does not change will agent in "thinking"

Problem Solving Agents Step 4 and 5

Execution phase, execute recommended actions Step 5: find a new goal and repeat

Ideal Rational Agent

For each possible precept sequence, do whatever action is expected to maximize its performance measure, using evidence provided by the percept sequence and any built in knowledge Also do actions in correct oder

Problem-Solving Agents Step 1

Formulate goal,

Problem-Solving Agents step 2

Formulate problem, decide what actions and states to consider given a goal. Decide how to quantify best solution Find states, operators and best solution

What is an uniformed search

Given no information about problem (Other than its definition)

What is an informed search

Given some idea of where to look for solutions

Measuring Performance: Time complexity

How long does it take to find a solution?

Measuring Performance: Space Complexity

How much memory is needed to perform the search?

Human agent sensors and effectors

Humans - Sensors: eyes, ears, etc. - Effectors: hands, legs, mouth, etc

Deterministic vs Stochastic

If next state of environment is completely determined by current state and the action executed by the agent. Can't predict environment in stochastic

Non-autonomous behavior

If no use of percepts then system has no autonomy.

Fully observable vs partially observable

If sensors give access to complete state of environment

Problem State Space

Initial State and Actions/Operators

Problem is defined by what four items

Initial state, Actions/Operators, Goal test, and path cost

Measuring Performance: Optimality

Is it the solution with the lowest path cost

Romania Example Environment

It is static: formulating and solving problem in a fixed environment Fully observable: all states knowable Discrete: cities are nodes and actions are links Deterministic: no randomness assumed Agent: single

Rational Agent

One that does the right thing ( to be most successful)

PEAS Description

Performance Measure Environment Actuators Sensors

Environment (automated taxi driver)

Roads, other traffic, pedestrians, weather, customers

Performance measure (Automated taxi driver)

Safe, fast, obey laws, reaach destination, comfrotable trip, maximize profits

How to search for solutions

Search through state space using search tree Search node, initial state( test if already goal state) Expand current state(Apply successor function to generate new states) Use SEARCH STRATEGY to determine which state to examine and expand next

Problem - Solving Agents Step 3

Search, Process of looking for best action sequence to reach goal

Romania Example, Search

Sequence of cities to get from Arad to Bucharest

Episodic vs sequential

Sequential environments require memory of past actions to determine the next best action. Episodic environments are a series of one-shot actions and only the current percept is relevant. Experience divided into atomic episodes. Next episode does not depend on previous episodes

Single agent vs multi agent

Solving a puzzle is single agent Chess is competitive multi-agent environment

Agent function

Specifying which action to take in response to any given percept sequence. that an agent's behavior is AGENT FUNCTION described by the agent function that maps any given percept sequence to an action

Romania Example, Formulate Problem

States: various cities Operators: drive between cities, Best solution: shortest time

Actuators (Automated taxi driver)

Steering, accelerator, brake, signal, horn, speak, display

Task Environments

The "problems" to which the ration agents are the "solutions"

Rationality depends on

The performance measure that defines degree of success Everything the agent has perceived so far(the percept sequence) What the agent knows about the environment The actions that the agent can perform

What is the job of AI

To design Agent Programs, though much current emphasis is on embodiment

Rationality is concered with what

With expected success given what has been perceived

Agent

anything that perceives its environment through sensors and acts upon that environment through effectors.