Intelligent agents

boorishadamantΤεχνίτη Νοημοσύνη και Ρομποτική

29 Οκτ 2013 (πριν από 3 χρόνια και 7 μήνες)

51 εμφανίσεις

560 Summer 2001

Intelligent Agents

Chad Tuneberg

Manish Ruwali

Robert Maulding

The authors of this presentation were MBA students of Southern Illinois
University while Dr. Eom was a visiting professor during the summer of 2001
academic year.

Why use an intelligent agent

A large pool of information to look into .

• Typical search engines may yield irrelevant results.

• Search engines are time consuming inefficient

and causes a lot of data traffic.

Intelligence is the degree of reasoning and learned behavior: The agent's ability to

carry out the task delegated to it.

What is intelligence ?

Intelligent agent

A software which performs a given task using information
from its environment to act in a suitable manner so as to
complete the task successfully. The software should be able to
adapt itself based on changes occurring in its environment, so
that a change in circumstances will still yield the intended
G.W. Lecky

Features of an Intelligent Agent

Social ability
: Agents operate without the direct
intervention of humans but interact with other agents and
sometimes humans via some kind of agent communication

: The ability of an agent to move around an
electronic network.


The agent updates its knowledge base every time
the user responds to the agent's suggestions and questions.

•Goal Oriented:

The assumption that an agent will act in order to
achieve its goals. An agent is capable of handling complex, high
tasks. The decision how such a task is best split up in smaller sub
tasks, and in which order and in which way these sub
tasks should be
best performed, should be made by the agent itself.


Do not simply act in response to their
environment, they are able to exhibit goal
behavior by taking the initiative.

•Collaborative behavior

Agents should be able to work in
concert with other agents, possibly via an agent
communication language, to achieve a common goal. Agents
may share knowledge and learning experiences in the process.

How do they work

They use a combination of statistical operations, artificial
intelligence, machine learning, inference, neural networks,
and information technologies.

Based Reasoning Techniques

:Agents use a
set of rules to decide which action or actions they
should take.These are specified by the user and the
agents builds on them to reach a solution.

Example. If rule A is satisfied then do action B
production rules in AI)
IBM’s RAISE (Reusable Agent Intelligence
Software Environment) is an example of rule
based reasoning.
It can
perform information flow functions: finding, searching, filtering,
categorizing, storing, routing, and/or selectively disseminating
information items.

Based Reasoning Techniques

Knowledge bases are built based on a specific
subject area or domain.
These then serve as the basis for
some inference mechanisms, including the rule
reasoning techniques.

This reasoning could be Case based or Machine learning
techniques Rule induction, Conceptual clustering,
Neural networks, Genetic algorithms.

This technique has been used by Cycorp, Inc., based in
Austin, Texas and is known as the Cyc knowledge base.

•Simple Statistical Analysis


It can determine the correlation among events of interest
using this information to give meaningful results.

Example This may help analyze and scan user actions to
find repeated sequence of actions.

Neural network:

This is a parallel distributed processor that has a natural
propensity for storing experiential knowledge and making it
available for use.
Neural networks handle unstructured
data .

Agent Communication Language


:Knowledge Query and Manipulation Language is
the language used for an application program to interact
with an intelligent agent.

This allows permissible procedures that allows agents to
interact with each other.

Provides the basic architecture for knowledge sharing by
acting as a communication facilitator.

The figure is based on the article "Towards Anticipatory Agents" by Ekdahl, Astor & Davidsson