UNDERSTAND MACHINE LEARNING IN JUST 10 MINUTES
Machine learning is nowadays getting
more popular and popular. it's a great deal, to be honest. There are limitless
possibilities using Machine learning opening doors to new dimensions. Let's get
started!!!
Basically, Machine learning is
training the machine (here computer) to do so and so tasks.
ML is all about collecting a hell lot
of data about something and feeding it to the computer. Don't get confused with the word machine, it's simply the computer itself. I will explain
things more elaborately by using examples.
Consider this situation,
I have the data of places( addresses)
that a person visits more frequently.
My data contains places like A, B, C, D
and E
Now by analyzing which place that
person visits most often, I could easily conclude that place as his or her
home.
Suppose out of the above 5 places, by
counting in an Excel spreadsheet,
One can get the count of how many
times that person visits this place in a day.
Count result
A - 7times
B- 3 times
C-2 times
D-1time
E-1time
So I can assume that A is that
person's home.
And B might be that person's College
or Office.
C might be his favorite coffee shop
or a restaurant.
See I am drawing assumptions from
just address data.
This is the same way Machine assumes
data and provide results. This service of getting data is really important
for Business Promotion. Because you know if that person visits the coffee shop
regularly, he is a coffee lover, and showing him coffee-related ads and stuff
like that will work! That's how even you get personalized feeds on the internet
based on your browsing pattern, this is how Machine Learning works.
It draws a conclusion from lots of data,
but the accuracy is never 100%.
The Machine makes false
assumptions too. In the above example itself, you can take place B as a college or
an office, so you can't get exact results.
Now let me explain to you with another
example.
Google has created an app for you guys to
understand how ML works, you guys can train this machine by providing faces
from the pictures and test the machine to find out if it could differentiate
between faces and other pictures. Try it again until the machine shows an
accuracy of 100%.
FREE ONLINE MACHINE LEARNING APP
This is a typical ML for facerecognition.
So now you know training the machine
is really hard when it comes to something with a lot of data. Example
identification of pictures of birds, for this you will have to train machine
using hell lot of pictures about every bird species on the planet, only then
will the machine identifies the correct bird images. Suppose you missed out on some birds like ostrich, then the machine won't identify ostrich as a bird, so know
you know, the machine requires millions of data to improve accuracy.
Otherwise, it can give false
assumptions.
You won't believe, nowadays, this ML
software is used even for CV screening so that the machine screens only this
Cvs with keywords mentioned in the job description, in this way even the best
candidates who missed out on some keywords will not be screened in. So better be
careful to read the job description and include the keywords in your CV,
because your CV is being read by the computer lol.
This is the essence of ML,
basically collecting lots and lots of data and training the Machine like you
train your young baby to differentiate among things and stuff.
Technically speaking, among the
collected data, it's divided into
1) Training data set
2) Test data set
All the data collected in the
training data set is tested or matched with the test data set to draw accurate
assumptions.
Here fairness is a big problem,
sometimes results won't be fair like the CV example.
In some CV's interests include caring
for the needy, family person, etc.
These CV's won't be screened in, which basically means this ML for CV screening is kind of biased.
So how can this problem of fairness
be solved??
The only thing that can be done to
improve the situation is by using tons and tons of diverse data.
Still, in the 21st century, the machine's accuracy is not 100%.
But you can really expect things to
get really better against all odds in the future.
There are endless opportunities in
different fields using ML, even for school admission or college
admission, ML can be used. The applications will be screened
through ML and only the screened candidates get an Interview Call for
Admission.
So systematic and fair training of
the Machine is really important, otherwise, it will give biased results.
All Marketing decisions are nowadays
made using ML. For example, if you searched Amazon for an iPhone, then the machine
will keep track of your search results and show you ads related to iPhones in
Google or other browsers. Nowadays you will even get an ad about iPhones in
Instagram and Facebook too, that's because they have linked it all using ML.
So ML has infinite opportunities as I
have told before, it's like giving training to Machine to make the Machine
think more advanced than the human brain.
In the army, this ML software can
be used to train robots to kill like Jarvis. Yeah, the Iron Man Suit.
Everything's possible by using fair and advanced training.
If we humankind misuses this amazing
technology, then we will bring a possible Human Robo War, yeah that's also very
possible.
The movie iRobot and Terminator and
other stuff is also possible.
If at all a machine is trained to
kill and act without emotions, then it can turn against the entire human race.
Whatever guys now I believe you get
something about how ML works and what are the advantages and problems
associated with ML.
Share your feelings in the comments
below.
Cheers
Subscribe and follow me
Check out the other posts in this blog
Check out the other posts in this blog
- MECHANICAL ENGINEER TOP PROFESSIONAL CERTIFICATIONS
- WILL CORONA GO BEYOND 2021? FIND OUT
- GOOGLE CERTIFICATION COURSES COMPLETELY FREE : ENROLL NOW
- UFC 249 FINALLY FIXED ON MAY 10TH SAT
- TOP 10 HAIR CUTS FOR TEEN BOYS AND MEN
- SSC JE MECHANICAL PREVIOUS YEAR QUESTION PAPERS PDF
- iPATE EXAM :THE GATE-WAY TO CORPORATE WORLD
- SSC JUNIOR ENGINEER EXAMINATION 2020 UPDATES - CUTOFF ANALYSIS, EXAM PATTERN & VACANCY
- FIVE TOP PAID ATHLETES IN THE WORLD
- THE LEGEND GOLDBERG
- ALL ABOUT APRIL 3RD, THE FUTURE PAST
- ALL ABOUT APRIL 2ND,THE FUTURE PAST
- REALLY USEFUL WHATSAPP TRICKS AND HACKS!!!
- My All time Five most gorgeous WWE Diva Superstars
- My Gym Experience and What is it like to have a six pack?
- FIVE WAYS TO MAKE WEALTH APART FROM STOCK MARKET
- WHICH SOCIAL MEDIA IS THE BEST?
- THE WRESTLEMANIA 36 THRILLERS
- THE APPEX PREDATOR VS THE RATED R SUPERSTAR
- THE PHENOM VS THE PHENOMENAL ONE
- BROCK LESNER SHOWS SIGN OF FEAR FOR THE CLAYMORE
Comments
Post a Comment