Mansoura University
Faculty of Computers and Information
Department of Computer Science
First Semester: 2020-2021
[CS324P ] Artificial Intelligence - 1 : INTELLIGENT AGENTS
Grade: Third Year (Computer Science)
Ass. Prof. Taher Hamza
Dr. Sara El-Metwally
Faculty of Computers and Information,
Mansoura University,
Egypt.
Contents
1 Intelligent Agents
2 Agent design
3 Environment Properties
4 Agents Types
2
Intelligent Agents
An agent is anything that can be viewed as perceiving its
environment through sensors and acting upon that
environment through actuators
Image Credit: Artificial Intelligence A Modern Approach Second Edition by Stuart J. Russell and Peter Norvig by 2003.
3
Intelligent Agents
Image Credit: Artificial Intelligence A Modern Approach Second Edition by Stuart J. Russell and Peter Norvig by 2010.
4
Intelligent Agents, examples
Human Agent
Sensors Eyes, ears, nose, skin,..
Actuators Hands, legs, mouth,..
Robotic Agent
Sensors Cameras, infrared ,…
Actuators Various motors, wheels,..
A software Agent
Keystrokes, file contents, received network
Sensors
packages…
displaying on the screen, writing files,
Actuators
sending network packets,…
5
Intelligent Agents
An agent’s behavior is described by the agent function which maps
from percept histories to actions:
f: P* A
Agent function will be implemented by an agent program which runs
on the physical architecture to produce f
agent = architecture + program
6
Intelligent Agents, example
Vacuum-cleaner agent
Percepts: location and contents, e.g: [A,Dirty]
Actions: Left, Right, Suck
Image Credit: Artificial Intelligence A Modern Approach Second Edition by Stuart J. Russell and Peter Norvig by 2003.
7
Intelligent Agents
Agent function as look up Table:
Image Credit: Artificial Intelligence A Modern Approach Second Edition by Stuart J. Russell and Peter Norvig by 2003.
8
Intelligent Agents
Agent function as look up Table:
An agent actions is completely specified by the
lookup table
Drawbacks:
Huge table
Take a long time to build the table
No autonomy
9
Intelligent Agents
Rational Agent:
For each possible percept sequence, Ideal rational
agent should do whatever action expected to maximize
performance measure, on the basis of built-in
knowledge agent has
Image Credit: Artificial Intelligence A Modern Approach Second Edition by Stuart J. Russell and Peter Norvig by 2003.
10
Intelligent Agents
Rational Vs. Perfection !
Omniscience is the unlimited knowledge
An omniscient agent knows the actual outcome of its
actions, and can act accordingly
Omniscience is impossible in reality
11
Agent design
Agent Design (PEAS)
Performance: How agent be assessed?
Environment: What elements exists around agent?
Actuators: How agent change the environment?
Sensors: How agent sense the environment?
12
Agent design
Agent Design (PEAS)
Automated taxi driver
Performance: Safe, fast, legal, comfortable trip, profits
Environment: Roads, other traffic, pedestrians, customers
Actuators: Steering wheel, accelerator, brake, signal, horn
Sensors: Cameras, speedometer, GPS, engine sensors, keyboard
13
Agent design
Agent Design (PEAS)
Part-picking robot
Performance: Percentage of parts in correct bins, speed
Environment: Conveyor belt with parts, bins
Actuators: Jointed arm and hand
Sensors: Camera, joint angle sensors
14
Agent design
Agent Design (PEAS)
Medical diagnosis system
Performance: Healthy patient, minimize costs, lawsuits
Environment: Patient, hospital, staff,…..
staff
Actuators: Screen display (questions, tests, diagnoses,
treatments, referrals)
Sensors: Keyboard (entry of symptoms, patient's answers)
15
Agent design
Assignment (1)
Design the agent for your projects?
(PEAS)
16
Environment Properties
Environment Properties (ODESDA)
Observable (or, partially observable)
An agent's sensors give it access to the complete state of
the environment at each point in time
Deterministic (or, stochastic)
The next state of the environment is completely determined by
the current state and the action executed by the agent
Episodic (or, sequential)
The agent's experience is divided into episodes, in each episode
the agent receives a percept and then performs a single action
17
Environment Properties
Environment Properties (ODESDA)
Static (or, Dynamic)
The environment is unchanged while an agent is deliberating
Discrete (or, Continuous)
A limited number of distinct, clearly defined percepts and actions.
Agent (single/ multi) (cooperative /competitive)
Number of agent in the environment
18
Note
https://www.geeksforgeeks.org/8-puzzle-problem-using-branch-and-bound/ https://www.fool.com/investing/what-does-the-future-hold-for-self-driving-cars.aspx
19
Environment Properties
Environment Properties (ODESDA)
The environment type largely determines the agent design
The real world is: partially observable, stochastic, sequential,
dynamic, continuous, multi-agent
Image Credit: Artificial Intelligence A Modern Approach Second Edition by Stuart J. Russell and Peter Norvig by 2010.
20
Environment Properties
Assignment (2)
Specify the agent’s environment for your project?
(ODESDA)
21
Agents Types
Agent Basic Types:
Simple reflex agents
Model-based reflex agents
Goal-based agents
Utility-based agents
22
Agents Types
Simple reflex agents
Choose actions only based on the current percept
Ignore the precept history (no memory)
Use condition-action rule
Very simple !
Image Credit: Artificial Intelligence A Modern Approach Second Edition by Stuart J. Russell and Peter Norvig by 2010.
23
Agents Types
Simple reflex agents: (dis-advantage)
The agent will work only if the correct decision can be made on the
basis of the current percept that is only if the environment is fully
observable
Infinite loops are often unavoidable – escape could be possible by
randomizing
Image Credit: Artificial Intelligence A Modern Approach Second Edition by Stuart J. Russell and Peter Norvig by 2003.
24
Agents Types
Model-based reflex agents
Action depend on history or unperceived aspects of the world
Need to maintain internal world model (state)
Without clear goal it is unclear to know what to do!
Image Credit: Artificial Intelligence A Modern Approach Second Edition by Stuart J. Russell and Peter Norvig by 2010.
25
Agents Types
Goal-based agents
Agents of this kind take future events into consideration
Agent has some goal information, choose actions according to goal
Some solutions to goal states are better than others!
What happed if we have conflicting goals!
Image Credit: Artificial Intelligence A Modern Approach Second Edition by Stuart J. Russell and Peter Norvig by 2010.
26
Agents Types
Utility-based agents
Try to Maximize agent expected happiness
Image Credit: Artificial Intelligence A Modern Approach Second Edition by Stuart J. Russell and Peter Norvig by 2010.
27
Learning agent
Image Credit: Artificial Intelligence A Modern Approach Second Edition by Stuart J. Russell and Peter Norvig by 2010.
28
Agents Types, example
Consider a chess playing agent, What sort of
agent would it need to be?
Simple-reflex agent: If yes? but some actions require some
memory (e.g. castling in chess)
Model-based reflex agent: If yes? but needs to reason about
future
Goal-based agent: If yes? but what about confliction
goals?
Utility-based agent:
Might consider multiple goals
29
Agents Types
Assignment (3)
Describe the agent type for your project?
Simple-reflex agent? Why? Why not?
Model-based agent? Why? Why not?
Goal-based agent? Why? Why not?
utility-based agent? Why? Why not?
30
31
