Let’s try to understand machine learning with examples. However, we have already learned that for machine learning what we should do.
Right?
In our previous section we discussed how through machine learning we train machines to perform various actions.
Then how do we define machine learning?
That’s the question a beginner will always ask. In a simple world, we can say that we train a machine to learn from data and do something.
After learning from data a machine may predict something. And the prediction should be close to accuracy.
Will a machine then act like an astrologer? 😀
Oh, no! 😡
It will act on scientific values, data, and do some tasks for us that we cannot do like a machine.
Therefore machine learning is a subset of science and we can say that it is a subset of computer science.
We program a computer in a way, so that it can learn from data.
Therefore, machine learning makes a computer or a machine able to learn something.
But we need to program a machine to do that.
What is the best example?
Well, the spam filter 👏 that our email service uses. Otherwise spammers would have bombarded our behind with hundred and thousand spam everyday. 😡
With reference to another very common example is face recognition.
We cannot just use the hardcoded approach to detect faces.
Why not?
Oh, our parents and grandparents have missed the opportunity to detect a beautiful face from others.
Because it remained an unsolved murder mystery until 2001. 🎉
An unsolved murder mystery is not there anymore.
Now every smartphone can detect faces. Even after learning a little you can also make an app that will detect faces. 😂
But why it remained an unsolved murder mystery?
Because a computer could not learn the way it does today.
Firstly, we know that pixels make an image in a computer. Right?
We humans don’t use pixels. 👍
On the contrary, we see images in a different way.
Whatever image it is.
Moreover, we don’t care about the rules that nature has made for us to perceive an image.
And here lies the biggest difference.
Computer knows only the rules. And we must train them to follow those rules so that they can learn from a big set of data and act almost like a human.
If you are a complete beginner your journey to learn TensorFlow might start from here.
For the TensorFlow beginners we have a dedicated category – TensorFlow for Beginners.
But besides that, you may need to learn several other machine learning and data science libraries.
As a result, you may check these categories as well – NumPy, Pandas, Matplotlib.
However, without learning Python, you cannot learn the usages of these libraries. Why? Because they all use Python as the Programming language.
Therefore please learn Python at the very beginning and start learning TensorFlow.
And, finally please check our Mathematics, Discrete Mathematics and Data Structures categories specially. We have tried to discuss from basic to intermediate level so that you can pick up the core ideas of TensorFlow.
Since images don’t represent themselves in pixels to us, we cannot define a good set of rules to detect one face from other digital images.
Therefore we need characteristics of images to set up some rules or an algorithm based on which a computer can detect faces.
This algorithm helps a computer to learn from a big chunk of data and do some tasks that a human cannot do.
This is machine learning.
Think about spam filters.
It is nothing but a machine learning program.
How does it learn?
It learns from the users. Because we flag emails and move them to spam folders.
The machine learning algorithm collects the data and learns from it.
There are more to come in this series.
We will discuss them one after another.
So stay tuned.
Leave a Reply