1.1 the “Turing test”. Which was able to

1.1 General Introduction


Machine Learning is
a sub field of artificial intelligence (AI). Machine learning is the
field of computer science that has ability to learn by itself without being
programmed clearly or in detailed way. Machine learning is also a field that is
growing significantly now days. In past computer algorithms used to be explicitly
programmed which was used by the computer to solve or calculate the given
problem. But now day’s machine learning algorithms instead allows computer to
train on data to improve its preference or solution for the given task.

Simply, the goal of machine learning is to analyze the
certain given data to a computer and improve the given task based on past
experience that can be understood and utilized by people.

 In 1950 a man named
Alan turning created the “Turing test”. Which was able to determine if computer
has a real intelligence. In 1952 for the very first time computer learning
program was written by Arthur Samuel and the name of the program was the game
of checkers. In this program the more user used to play the game computer used
to improve at the game by studying which moves made up winning strategies and
incorporating those moves into its program

1.2 Current scenariooverview.


Machine learning is becoming a dominant field of a computer
science so, machine learning is the future of computer science which must be
learned by today’s generation people in order to bring dramatic revolution in
field of computer science. As this is the age of big data machine learning is
being used in various field of science, from astronomy to biology as well as in
everyday life of people, as we use digital devices more data is continuously
being generated and collected as well. Those data may not be of any use to many
people but, some smart people finds a new ways to use that data and turns it
into a useful product or service. In this transformation machine learning plays
a huge role.

2.1 Elaboration.

In today’s world Machine
learning has changed the way that technology used to perform given task. For
example, let us consider a supermarket that has huge showroom for all kinds of
goods. Those goods are sold to millions of customers all around the world. So,
every day there is a huge transaction that are stored in computer. In
supermarket customer wants to find the goods in cozy way that suits them or
their work and that satisfy their needs. Whereas owner of the supermarket wants
to increase the profit and sales of the goods by predicting customers need and
demand which is about next to impossible without machine learning. So, to solve
this problem we need an algorithm to run in computer, which we don’t have. But,
supermarket have data of every customer like what customers were looking for,
what they bought. Analyzing such data helps us understand the process and we
can predict what customer will buy or interested in that helps owner to
maximize the sales and profit as well. (Alpaydin, 2014/8/22)

There are some of the
real-world application of machine learning that are already used in real life
they are:

Speech recognition:

Now days Speech recognition is in more practice
then before. Speech recognition enables the recognition of spoken language into
text form by computers, which uses machine learning in order to train the
system to recognize speech. Because there is high rate of accurate result when
system is trained rather than untrained system.


Computer vision:

Some of the computer vision that are developed
by using machine learning are face recognition system and system that classify
microscope images of cells automatically. For example in us more than 85% of
the hand written mails are arranged automatically, using trained software that
uses machine learning.



Machine learning is playing very important role
in detection the diseases. For example, the project called RODS collects the
data of admission reports to emergency rooms across western Pennsylvania, and
with the use of machine learning software the data of admitted patients are
analyzed in order to detect the symptoms for a particular patients diseases and
their geographical distribution. Some current work involves adding of data of
purchased medicine in medical stores to improve the machine learning system.


Robot control:

Machine learning is wildly used in robots
specially to acquire control strategies. For example there was a completion
called Darpa-sponsored that involved 100 miles running race in the desert which
was won by a Robert that used machine learning in which Robert self-collected
the data and used it in detecting the distance objects due to which Robert was
able to win.


Accelerating empirical sciences:

Machine learning is changing the way of many data-intensive
empirical sciences. For example machine learning is used to analyze the gene of
particular person to discover unusual astronomical objects by collecting
massive data by the Sloan sky survey to characterize the complex patterns of
brain activation that indicate different cognitive states of people in FMRI

3.3 Technical skills of
machine learning:


To implement machine learning we require good
knowledge of java/python/c++. Each of this programming language has its own
role to play in machine learning. Java helps in compiling/debugging, c++ helps
in speeding cod up and python contains machine learning algorithm that produce
compact. And all of these programming language courses are available in Nepal.

Applied math and algorithms:

Math and algorithms play very important role in
machine learning without it machine learning cannot function. Because in order
to function machine learning need curtain algorithms and math that helps to
understand subject and discriminate models.


Distributed computing:

Machine learning takes huge number of data sets
in order to perform task as it has to analyze past data and improve the
algorithms on its own. Storing huge number of data in a single machine is not
possible so we need different computer in order to process data.


Learning more about advance signal processing

There are lots of signal processing techniques
now days some of them are contour lets, shearlets, bandlets which can be used
to solve our problems. (strife, 2015)


3.4 Hardware requirement:

Graphics processing unit(GPU):

We require a good Graphics processing unit in
order to perform given task smoothly. Some of the Graphics processing unit are GeForce
9000series, GeForce 10series etc.

Central processing unit(CPU):

In order to run machine learning algorithm we
require a high central processing unit like 3.8 GHz with core i7-6850k. If we
require highest cup then we there is core i9 as well which is latest one.

System memory:

At least 8 GB of memory is needed which can be
changed later up to 64 GB later in motherboard.


If we have SSH hard disk it is good because it
is faster than HHD but if we have HHD one there should be lots of space. For HHD
at list 1TB is required.


Cooling computers helps to maintain the
temperature of computer as computer gets heated during its long time use. If
computer get heated more it will effect in its preformation so, cooling is

Power supply:

If require high capacity of computers for machine learning
so, high power supply is also needed form 1400 to 1600 watts