Topic 2 - Applications of AI

clangedbivalveAI and Robotics

Oct 19, 2013 (3 years and 9 months ago)

68 views

Topic 2
-

Applications of AI

St Kentigern’s Academy

Unit 2


Artificial Intelligence

Artificial Neural
Network/System


A
neural network

is an electronic model
of the brain consisting of many
interconnected simple processors.

Artificial Neural
Network/System


Neural networks take a different
approach from conventional
programming. Neural networks are
designed to try and do things that we
don’t exactly know how to do.

Artificial Neural
Network/System


The brain consists of millions of
interconnected records.



An ANS consists of hundreds of
interconnected artificial neurons, so it
is based on the same model as the brain.

Artificial Neural
Network/System


Applications of ANS


Debt risk assessment


Loan granting is one area in which neural
networks can aid humans, as it is not based on
predetermined criteria, but instead the answers
are vague.



The process works by analysing past failures
and making decisions based upon past
experience.

Artificial Neural
Network/System


Applications of ANS


Stock Market Prediction


Neural systems have been touted as all powerful
tools in stock market prediction.



The program tries to anticipate the market by
interpreting external parameters such as
economic indicators and public opinion.

Artificial Neural
Network/System


Applications of ANS


Postal Services


Character recognition is applied to handwritten
documents and is used to help post offices
automate the sorting of mail.

Artificial Neural
Network/System


Advantages


They learn and adjust to changes in
circumstances.



Disadvantages


The neural system need to be trained which
is very time consuming.

Vision Systems


This is the machines ability to make sense
of visual input.



A camera sends its image to a computer


The computer compares the images of items
with a ‘learned’ image of an ideal item of the
same type.



Some vision systems use an ANS to help
with image interpretation.

Vision Systems


Applications of Vision Systems


Industrial


Vision systems are used to check items on production lines



Military


Target recognition is a military application which uses image
data to determine if an enemy target is present.



Satellite Photo Interpretation


Satellites collect an enormous amount of data which would
normally take researchers ages to analyse and interpret.
Using a vision system with a neural network speeds up the
task.

Speech recognition



These systems allow the user to
communicate with the computer by talking
to it.

Speech recognition


The characteristics of speech recognition
are:


Training for each voice pattern;



Control instructions;



Background noise;



Factors affecting accuracy.

Speech recognition


Training



However, before using the system you must
first go through the process of
training

it.



This consists of reading pre
-
defined text into the
computer.



The computer samples you voice and matches it to
sounds which are common in all words.

Speech recognition


Training



When training a speech recognition package you
should take the following into account:



Speak in a consistent, level tone;



Use a consistent rate, without speeding up;



Speak without pausing between words;



Work in a quiet environment;



Pronounce words clearly.

Speech recognition


Control Instructions


As well as allowing a user to speak into a
system and it converting the sound to text,
systems also allow a user to speak commands.



This is seen as a different form of HCI as it
is possible to control the computer without
using a keyboard or a mouse. The user can
switch from dictation mode to command mode
easily when working on a document.

Speech recognition


Background Noise



A change in the background noise will affect
the accuracy of the system.



A system that is trained in an empty quiet
room will make mistakes if taken and used in a
room with background noise, e.g. a room full of
computers will make a buzzing noise that a
system will pick up on.

Speech recognition


Factors Affecting Accuracy



These include:



Background noise;


The user having a cold/sore throat;


The user’s accent/dialect;

Speech recognition


Applications of Speech recognition



Word processing


The user speaks the words and they are converted
into text and put in a WP document;



Punctuation Commands


As another form of HCI to input commands to the
computer;



Disabled Users


For users who are unable to use their hands;

Speech recognition


Applications of Speech recognition



Cars


Some cars let you control mobile phones/music via
speech recognition.



Military


Used in training simulators where a speech
interface can simulate multiple remote operators;



Mobile Phones


Rather than trying to type using the small
keyboard you can tell the phone who to call.

handwriting recognition


The user writes on the touch
-
screen using a
stylus pen and the writing is turned into text
and displayed on the screen.



Early systems (and still some today) required
the user to enter letters and numbers so that
they can be stored and matched against any
future input.



Current systems train the user to write in a
specific way, this means that more users can use
the system as it is not programmed to one
persons handwriting.

handwriting recognition


Applications


Palmtops and Tablet PC’s


Both of these are small devices and by having
handwriting recognition available via the touch
-
screen, there is no need to use up space with a
keyboard.

Intelligent Robots


Robots are automatic electro
-
mechanical machines.



Some are not fully automatic but are remotely
controlled by humans using TV and other feedback.



The term intelligent robot has been coined to
differentiate robots equipped with vision systems,
ANS, the ability to learn or some other aspect of
AI.



A dumb robot has no controlling processor and it
merely repeats the same movements again and
again.

Intelligent Robots


Advantages of Intelligent Robots



Increased productivity


Robots do not require breaks



Improved accuracy


More accurate than humans



Consistency


They don’t get tired like humans



Reduced Wage Bill


Fewer staff employed with no wages

Intelligent Robots


Advantages of Intelligent Robots



Increased productivity


Robots do not require breaks



Improved accuracy


More accurate than humans



Consistency


They don’t get tired like humans



Reduced Wage Bill


Fewer staff employed with no wages



Hostile Environment


Can go to dangerous places where humans can’t

Intelligent Robots


Sensors


These allow the robot to be aware of their surroundings



Bumb Sensors


When the robot touches an item it knows;



Proximity Sensor


The robot knows when it is near something



Light Sensor


Can tell the differences in light



Temperature sensors


The robot can tell if the temperature has changed

Intelligent Robots


Applications


Automated Delivery


Large companies employ a team of staff to deliver internal mail
each day. Robots are now being used to automate the process.


Pipe Inspection


Robots are proving useful for routine pipe inspection in
manufacturing and processing plants. Their ability to reach
inaccessible areas makes them useful.


Bomb Disposal


Robots for this purpose are available in many communities and are
used routinely by law enforcement. Their objective is to
separate personnel from danger as they seek out and detonate
explosives.


Exploration of unknown environments


A robot which is being sent to another planet must have
independent decision making capabilities.

Expert Systems


Expert systems collect the small
fragments of human know
-
how into a
knowledge base which is used to reason
through a problem, using the knowledge
which is appropriate.



Instead of attempting to create an
intelligent program, research focused on
creating a means of representing and
accessing knowledge.

Expert Systems


Expert systems are computer programs
which could offer advice in a restricted
subject where it was possible to create
facts and rules representing knowledge.



An expert system is an attempt to replace
the human expert and to make their
knowledge available in a cost
-
effective form.

Expert Systems


Advantages of Expert Systems



More available than a human expert



In a remote area an expert may not be available. The expert
system could be available 24/7 and consulted at any time.



Don’t have to pay the system wages



Human experts can command large fees for their services, but
once an expert system is set up the company wage bill can be
reduced by employing fewer people.


Expert Systems


Advantages of Expert Systems



They have the combined knowledge from more than 1
human expert



A single human has only his knowledge. An expert system can
contain the combined knowledge of many experts in the field.



They are more reliable.



Humans are prone to error. Expert systems have a
restricted
domain
and are only aware to the task for which they were
designed and are not prone to interference from external
factors. This makes them more reliable.

Expert Systems


Applications


Medicine


MYCIN was the first expert system which gave advice to doctors
on blood disease.



DHSS


An expert system advices on the benefits due depending on
circumstances.



Legal


Solicitors can check that they are giving the right advice to a
client



British Gas


Has an expert system which is used to calculate the most likely
place where corrosion will occur in a gas pipe.

Expert Systems


Social, Legal and Ethical Issues


Effects on employment



Many fear that using AI in industry will mean a
decrease in jobs and with good reason. On
-
line
shopping means fewer sales clerks.



However, we need expert systems because of the lack
of experts.



And remember, an expert system should only be used
as advice. Therefore a human needs to query the
expert system to obtain the knowledge. This means
that there are training implications.

Expert Systems


Social, Legal and Ethical Issues


Training Issues


Staff will need to be trained to use the new system.



Some staff will be resistant to change and worried
about job losses. Staff need to be reassured that an
expert system will help them and that they will be give
all the training they need in order to do their job.

Expert Systems


Social, Legal and Ethical Issues


Public Reactions


An example: if you had to go to the doctor for a
serious consultation, would you prefer to see a human
or a robot?



Most people would not like to think that important
decisions are being made for them by a computer
program.

Expert Systems


Social, Legal and Ethical Issues


Loss of Human Expertise


There is a fear from experts that if they give their
knowledge to a program then they will be no longer
needed.



This is not the case. Expert systems are required
because there are not enough experts.

Expert Systems


Social, Legal and Ethical Issues


Who is to blame?


If an expert system gets it wrong and a human follows
that advice and uses it, who is to blame?


The program?


The programmer?


The Human Expert?



An expert system is to be used the same way that a
book should be used as reference. If the human
chooses to agree with the expert system that’s fine.
But if the advice is wrong it is the human who is
responsible.