AI Intelligent Agents

download AI Intelligent Agents

of 15

Transcript of AI Intelligent Agents

  • 8/10/2019 AI Intelligent Agents

    1/15

    PSUCS 370Artificial Intelligence Dr. Mohamed Tounsi

    Artificial Intelligence

    2. Intelligent Agents

    Dr. M. Tounsi

  • 8/10/2019 AI Intelligent Agents

    2/15

    PSUCS 370Artificial Intelligence Dr. Mohamed Tounsi

    Definition

    Is anything:

    perceiving its environment through sensors

    acting upon its environment through effectors

    Example:With Robotic agent

    Sensors: Cameras and infrared

    Effectors: various motors

  • 8/10/2019 AI Intelligent Agents

    3/15

    PSUCS 370Artificial Intelligence Dr. Mohamed Tounsi

  • 8/10/2019 AI Intelligent Agents

    4/15

    PSUCS 370Artificial Intelligence Dr. Mohamed Tounsi

    Definitions and Concepts

    Rational Agent One that does the right thing (most successful!)

    Issue:how and when to evaluate the agents success ?

    Performance Measure

    Criteria that determines how successful an agent is

    Percept Sequence

    Everything that the agent has perceived so far

    Ideal Rational Agent

    Should do whatever action is expected to maximize itsperformance measure based on percept sequence and whateverbuild-in knowledge the agent has

  • 8/10/2019 AI Intelligent Agents

    5/15

    PSUCS 370Artificial Intelligence Dr. Mohamed Tounsi

    Rational agents

    An agent should strive to "do the right thing", based onwhat it can perceive and the actions it can perform. Theright action is the one that will cause the agent to be mostsuccessful

    Performance measure: An objective criterion forsuccess of an agent's behavior

    E.g., performance measure of a vacuum-cleaner agentcould be amount of dirt cleaned up, amount of time taken,amount of electricity consumed, amount of noisegenerated, etc.

  • 8/10/2019 AI Intelligent Agents

    6/15

    PSUCS 370Artificial Intelligence Dr. Mohamed Tounsi

    PEAS

    PEAS: Performance measure, Environment,Actuators, Sensors

    Definition: Actuator = Effector

    Must first specify the setting for intelligent agent design

    Example: the task of designing an automated taxi driver:

    Performance measure: Safe, fast, legal, comfortable trip,maximize profits

    Environment: Roads, other traffic, pedestrians, customers Actuators: Steering wheel, accelerator, brake, signal, horn

    Sensors: Cameras, sonar, speedometer, GPS, odometer,engine sensors, keyboard

  • 8/10/2019 AI Intelligent Agents

    7/15PSUCS 370Artificial Intelligence Dr. Mohamed Tounsi

    PEAS

    Agent: Medical diagnosis system

    Performance measure: Healthy patient,minimize costs,

    Environment: Patient, hospital, staff Actuators: Screen display (questions, tests,

    diagnoses, treatments, referrals)

    Sensors: Keyboard (entry of symptoms,

    findings, patient's answers)

  • 8/10/2019 AI Intelligent Agents

    8/15PSUCS 370Artificial Intelligence Dr. Mohamed Tounsi

    PEAS

    Agent: Part-picking robot

    Performance measure: Percentage of parts incorrect bins

    Environment: Conveyor belt with parts, bins Actuators: Jointed arm and hand

    Sensors: Camera, joint angle sensors

  • 8/10/2019 AI Intelligent Agents

    9/15PSUCS 370Artificial Intelligence Dr. Mohamed Tounsi

    PEAS

    Agent: Interactive English tutor

    Performance measure: Maximize student's scoreon test

    Environment: Set of students Actuators: Screen display (exercises,

    suggestions, corrections)

    Sensors: Keyboard

  • 8/10/2019 AI Intelligent Agents

    10/15PSUCS 370Artificial Intelligence Dr. Mohamed Tounsi

    Agent functions and programs

    An agent is represented by the agent functionwhich maps percept sequencesto actions

    Aim: find a way to implement the rational agentfunction concisely

  • 8/10/2019 AI Intelligent Agents

    11/15PSUCS 370Artificial Intelligence Dr. Mohamed Tounsi

    Table-lookup agent

    FunctionTable-Agent(percept) returnsan ActionStatic:percepts// sequence, initially empty

    Table // table of actions, indexed bypercept, initially fully specified

    Begin

    Appendperept to the end of the perceptsAction := LOOKUP(percepts, Table)

    returnAction

    End

    Drawbacks:

    Huge table

    Take a long time to build the table

    No autonomy

    Even with learning, need a long time to learnthe table entries

  • 8/10/2019 AI Intelligent Agents

    12/15PSUCS 370Artificial Intelligence Dr. Mohamed Tounsi

    Program Agent

    FunctionSKELETON-AGENT (percept)

    Static: memory;

    memory

  • 8/10/2019 AI Intelligent Agents

    13/15PSUCS 370Artificial Intelligence Dr. Mohamed Tounsi

    Agent types

    Four basic types in order of increasing generality:

    1. Simple reflex agents

    respond immediately to percepts (i.e.: rules)2. Model-based reflex agents

    3. Goal-based agentsact so that they will achieve their goal(s)

    4. Utility-based agentstry to maximize their own happiness

  • 8/10/2019 AI Intelligent Agents

    14/15PSUCS 370Artificial Intelligence Dr. Mohamed Tounsi

    Example: Simplex reflex agent

    FunctionReflex-Vacum_agent(location, status)return actions

    {

    If status = dirty then action= Clean_ON;else {

    if location = A then action=right;

    else action = left;

    }

    }

  • 8/10/2019 AI Intelligent Agents

    15/15PSUCS 370 Artificial Intelligence Dr Mohamed Tounsi

    If the agent is based completely on the built-inknowledge part, such that it pays no attention to itspercepts, then we say that the agent lacks autonomy

    If the behavior of the agent is determined by only its own

    experience the system is autonomous

    Autonomy

    AI Agent should have initial knowledge as well as anability to learn