Types of Agents

• Agent = Architecture + program

(A) Simple Reflex Agents:
SENSORS E
N
Agent V
What the
I
world is like R
now ? O
N
M
E
What action I
Condition Action N
should do now?
rules T

ACTUATORS

AI- RCS 702 Chahat Sharma (Asst. Prof.)

 Simple Reflex Agent

• Selects actions on the basis of current percept, ignoring rest of the

percept history.

• E.g. : Vacuum Agent makes decision based on its current location &
whether the floor is dirty or not.

• A Condition Action rule can be established.

AI- RCS 702 Chahat Sharma (Asst. Prof.)

 Simple Reflex Agent

• E.g : In automated driver vehicle if car in front brakes , and brakes

light comes on then , agent must perform some action like to put the
brake.

if (car – in front is braking) initiate braking.

Code for Simple Reflex Agent

function sim_reflx (percept) returns an action

static: rules, a set of condition – action rules // Static KB
state Interpret _ input (percept)
rule match(state, rules)
action Action[rule]
return action

AI- RCS 702 Chahat Sharma (Asst. Prof.)

 Simple Reflex Agent

Function TABLE-DRIVEN_AGENT(percept) returns an action

static: percepts, a sequence initially empty

table, a table of actions, indexed by percept sequence

append percept to the end of percepts

action  LOOKUP(percepts, table)
return action

AI- RCS 702 Chahat Sharma (Asst. Prof.)

 Model Based Reflex Agent

Sensor
state
What the
world is like
now ?
How the world
evolves ?

Environment
What my actions
do ?

Condition-action rules What action I

should do now

Agent Actuator

AI- RCS 702 Chahat Sharma (Asst. Prof.)

 Model Based Reflex Agent

• To tackle partially observable environments.

 Maintain internal state

• Over time update state using world knowledge

 How does the world change.
 How do actions affect world.

AI- RCS 702 Chahat Sharma (Asst. Prof.)

 Goal Based Agent

Sensor
state
What the
world is like
now
How the world evolves

Environment
What it will be
like if I do Action
A
What my actions do

What action I
should do now
Goals

Agent Actuator

AI- RCS 702 Chahat Sharma (Asst. Prof.)

 Goal Based Agent

• The agent needs a goal to know which situations are

desirable.

“ Things become difficult when long sequences of actions

are required to find the goal.”

• Typically investigated in search and planning research.

 Major difference: future is taken into account

 Is more flexible since knowledge is represented explicitly

and can be manipulated.

AI- RCS 702 Chahat Sharma (Asst. Prof.)

 Utility Based Agent

Sensor
state
What The world
is like now
How the world evolves

Environment
What it will be like
if I do Action A

What my actions do

How happy I would

be in such a state ?

Utility

What action I
should do now

Agent
Actuator

AI- RCS 702 Chahat Sharma (Asst. Prof.)

 Utility Based Agent

(i) Certain goals can be reached in different ways.

 Some are better, have a higher utility.

(ii) Utility function maps a (sequence of) state(s) onto a real

number.

(iii) Improves on goals:

 Selecting between conflicting goals
 Select appropriately between several goals based on
likelihood of success.

AI- RCS 702 Chahat Sharma (Asst. Prof.)

 Summary of Last Lecture

• AI Terminologies
• Rational Agent
• PEAS
• Environment Types

Today’s Topics of Lecture

• Learning Agent
• Structure of an Agent
• Natural Language Processing
• Computer Vision

AI- RCS 702 Chahat Sharma (Asst. Prof.)

 Learning Agent

Stimuli Feedback
Environment examples Learner
or Component
Problem
Generator

Knowledge Critic
Base Performance
Evaluator

Response
Performance
component

Tasks

AI- RCS 702 Chahat Sharma (Asst. Prof.)

 Structure of an Agent

Steps:
• Define the task environment

• Define PEAS- Performance, Environment, Actuator and Sensor

• Define or tabulate agent function (i.e., percept sequence or action

column)

• Design agent program

• Design architecture to implement agent program

• Implement an agent program

AI- RCS 702 Chahat Sharma (Asst. Prof.)

 Natural Language Processing

 Computer languages are used to convert natural languages into

synthesized form in m/c.
 In order to understand natural language m/c must:
 Generate it.
 Understand
 Translate

Natural Knowledge NL text, c m/c

language Parser o/p
representation code,
Text translati
System Or language
string on
o/p

dictionary

AI- RCS 702 Chahat Sharma (Asst. Prof.)

 Components of Natural Language Processing

• Natural Language Understanding

• Natural Language Generation

Natural Language Understanding

• Speech Recognition
• Syntactic Analysis
• Rules of Syntax (Grammar is defined)
• Parsing (Form Tree)
• Semantic Analysis
• Pragmatics

AI- RCS 702 Chahat Sharma (Asst. Prof.)

 Natural Language Understanding

“The bat eats a cat” Parse Tree

Grammar

s np, vp
np d, n
vp v, np
dthe/a
nbat/cat
veats

AI- RCS 702 Chahat Sharma (Asst. Prof.)

 Natural Language Understanding

AI- RCS 702 Chahat Sharma (Asst. Prof.)

 Natural Language Generation

• Text/ Speech Planning

• Talking Back
• Medium to talk back/How to talk
• Speech Synthesis

AI- RCS 702 Chahat Sharma (Asst. Prof.)

 Computer Vision

“Transformation of data from a still or video camera into either a

decision or a new representation.”

I/P data : Contextual info. ..e.g: camera mounted in a car

Laser range finder indicates an object as 1 meter away.

Decision : Is there a person in this scene ?

Are there 14 tumor cells on this slide

New Representation: Turning a color image into a gray scale image ,

Removing camera motion from an image
sequence.

AI- RCS 702 Chahat Sharma (Asst. Prof.)

 Computer Vision

In m/c vision, computer receives a grid of numbers from the

camera or from disk………..

Data is corrupted by noise which can be :

 Variations in environment viz.. Weather, lighting , reflections,
movements
 Imperfect lens with low resolution
 Imperfect mechanical set up of imaging system
 Finite integration time on sensor like: Motion Blurring
 Electrical noise in the sensor/electronics
 Compression artifacts after image acquisition.

AI- RCS 702 Chahat Sharma (Asst. Prof.)