Artificial intelligence machine (AIM)
Using Internet
Ashok.G
Deportment of
Computer Applications, SNIST,
Abstract
A large amount data is
available in the internet and that data is increasing day by day. We know that
there is no effective use with raw data. So that acquiring knowledge from that
data is important and it is not so difficult to a man to gain knowledge by reading
the data that available in the internet but it is a time consuming task. To a
machine it is a challenge to gain knowledge from the internet but if it succeeds
it reduces the user time. And it is must an efficient machine to achieve this
and as well as it should having with a good architecture. This paper discus
about how to acquire the knowledge from the internet by a machine called
Artificial Intelligence Machine (AIM). And
architecture is proposed in this paper to gain the knowledge from the internet
by a machine.
Introduction
Today’s world has
several raw data in the internet and it is important to use that data
effectively. To gain knowledge from those data available in the internet is
time consuming to a man and it is difficult task to a machine. A very
intellectual machine is needed to gain knowledge from the internet and as well
as it must be useful to interact with the human being. To achieve this, a good
architecture is needed. This paper is attempted to give a proposal of
architecture of artificial intelligence machine (AIM). And this machine should
be gain knowledge from the internet and that knowledge should be stored in the
knowledge base. When the user needs information from the internet he can able
to go by the artificial intelligence machine. Then the machine gains the
knowledge from the internet and then gives it to the user in the suitable
output mode, the output may be in voice, in visual, in audio or in the
combination of all, which is comfortable to the user or if there it is must to
use the output component it is used by the machine. For example if the user
wants to see the image of a banana tree it will show on the screen. This
machine should have to give the correct information to the user without or with
less errors to achieve this, the proposed machine utilizes the features of data
mining and as well as the techniques of Artificial Intelligence (AI).
Architecture
of AIM
The following figure
illustrate about the Architecture of AIM, which contains several components
those placed in the diagram. Input receiver receives the information from the
user it allows the information in any type of format like audio, voice, text,
image etc. and that input received from the user is converted into the text
format this text format will be sent to Question Generator (QG). And question
generator generates all possible questions and it interacts with the search
engine and as well as with the AI section in the process of answering the
question. Checks for the answer firstly in the AI section and then proceeds
according to the result received from the AI section. Question Generator is an
important component in machine because effective search is an important task to
answer the question.
Artificial
intelligence machine Architecture
Question Generator
generates the questions by adding who, whom, where, when how, why etc.
according these questions the searching process will be done. Search engine
gets the questions from the question generator according these questions the
search will be done by the search engine. And the search engine searches about
the information in the internet. All the search result will be given to the
information repository. And that store the information temporarily. This
information will be analyzed by data mining engine. The useful information
gained by data mining engine is converted into facts and rules in the AI
section then these facts and rules will be stored into knowledge base. Wait
component stores the questions temporarily. If it is necessary it interacts
with AI program. AI program gives the result to the QG and as well as to the
output section. And output section gives the output to the user in selected
format.
Working
with AIM
Machine is designed to
accept the questions from the user in any format and this input is converted
into text format later. The input receiver receives the question in any
mode/format and then it converts into text mode as mentioned above. Let us
suppose a user asked a question to the machine. Who was the father of first
prime minister of India? In the text format the input receiver takes it as an
input and then it sends to Question Generator. If it is in another format converts
it into text format and sends it to QG.QG divides the input into all possible
questions in the example the questions will like,
i.
Who was first prime minister of India?
And from i we can get
X.
ii.
Who was father of X?
Here we can get two
cases one is discuss about how the answering procedure done by using already
existed facts and second case illustrate how answering to the questions by using the internet. In
the second information got from the internet is converted into facts and then
by using those facts answering is done.
CASE1:
Then first question
sent to the AI section here the question will converted into AI form
First_pm_of_India(X);
It is given to the AI
program as a goal and the program finds the solution to X. Here X is solved by
using existed facts and rules. X is sent to the Question Generator (QG). Then
QG is updates the second question. Then second question will become like below
if X=Jawaharlal Nehru.
Who was the father of
Jawaharlal Nehru?
QG sends the updated
second question to the AI section again and it will convert it into AI form in
the AI section. The converted form is like below.
Father
_of(Jawaharlal_Nehru,Y)
And then above fact is
send to the AI program as a goal and it tries to solve value of Y by using
existed facts and rules. If Y is solved the value of Y is return to output
section and then the output section converts the output into user required
format like audio or image. AI program sends the signal SUCCESS as a result.
But there is a possibility to find no solutions to X and Y. If the AI system
fails to solve the X it initially sends the result NORESULT to question
generator. Then QG goes for case 2,which
searches about the result in the internet. About case2 is explained in the
below. In this situation these two questions are stored in the wait section.
Storing
the two questions in the wait state
In the
second situation we got the solution to the first question but not for the Y.If
the AI system fails to solve Y result NORESULT to Y with solved X will send to
QG then QG enters case 2 with question. We know that value of X then QG
proceeds with X value to find value of Y
X=Jawaharlal_Nehru
We got it from the
AIprogram and QG enters to case 2 with second question like below.
Who was the father of
Jawaharlal Nehru?
In this situation only
second question will stored in the wait section
Storing
the second question in the wait state
CASE2:
If result is NORESULT from
the AI machine then Questions Generator sends the question to the search Engine
as keywords. In this section search engine performs searching in the internet.
The search engine searches in the World Wide Web and gives the search result to
information repository here all the search results will be stored temporarily. From
this information repository the useful information is retrieved by using data
mining techniques and it send to AI section. In this section all the useful
information gained from the mined data will be converted into facts and rules
these facts and rules are stored in the knowledge base.
For example we got the
information like below
Pundit Jawaharlal Nehru
was the first prime minister of India and he was also first central minister
for external affairs. His father was Mothilal Nehru and his daughter was Indira
priya dharshini.
The above information
is converted into facts like below
First_PM_of_India(Jawaharlal_Nehru).
First_Minister_of_Externalaffairs(Jawaharlal_Nehru).
Father_of(Jawaharlal_Nehru,Mothilal_Nehru).
Daughter_of(Jawaharlal_Nehru,Indira_Priyadharshini).
These facts will be
stored in the knowledge base and then tries to answer the questions in the wait
section If succeed to answer the questions result will be send to the output
section then there the output will converted into required format and gives to
the user. And sends success signal to the QG then it stops the process of
searching.
If the knowledge base fails to solve the
questions that are in the wait section then AI program sends signal FAIL to QG
and then repeats the search with different way and continues the process for
several times if require and if the time is expired to solve the question given
by the user sends EXPIRE signal to the QG and SORRY to the user.
Conclusion
To reduce the error
rate it still important to develop the architecture and it is important to
increase the efficiency of the system. Then only the artificial Intelligence
Machine becomes more users friendly. To make the system very efficient it is
must that system should generically answer to all kinds of questions. But it is
sometimes difficult to answer the questions related to science and mathematics
so that it is still a challenge. And sometimes the related information may not
available in the internet. Sometimes it may possible in the internet there is
false information and it is difficult to analyze the system to recognize about
false information.
