Understanding Machine Learning in Child Language
This is going to be fun but informative and helps novice learners to understand Machine Learning.
Firstly!!!… Don’t get intrigued about the word ‘Machine Learning’ because it is everywhere now…….
I would take an example and bring out the term ‘Machine Learning’. Hope, this will help everyone to understand this most fascinating term.
It’s Sunday!! and Family went out for lunch, after having lunch, they were on their way to the child’s grandmother's home. While walking on the road, a girl noticed an object passing next to them, which she was unaware of. This little girl was at the age of recognizing things.
In their grandmother’s home, Father and daughter are standing on the balcony and having a conversation in the following way :
By pointing out her finger at the object, she asked her father
Daughter: “What is that object that is parked there, which is laughing at us ?”
Father: It’s called “Car”, my dear.
By pointing out her finger again at another object, she asked her father
Daughter: “What is that object that is parked there, which is looking angry ?”
Father: It’s also called “Car”, my dear.
Father also explains that “Any object which is having four wheels, having similar dimensions can be considered a Car, irrespective of its color”.
Now, the father asks the daughter to recognize the object (which is in the below image)
Daughter: “It’s a Car”
Father: You are correct, Well done!! my dear.
What happened between father and daughter all this while?
It’s nothing but just ‘LEARNING’. The daughter is LEARNING from his father and the father is TEACHING his daughter (Training her by sharing his knowledge).
If you replace ‘Daughter’ with ‘MACHINE’ and ‘Father’ with ‘YOU’. It’s called ‘MACHINE LEARNING’.
When you TEACH the machine in the same way (about Car’s example, which is illustrated above) —This is called ‘MACHINE LEARNING’ (because here the machine is learning from you)
How will you make Machine LEARN?
Father and Daughter can communicate in a language (which can be English or their native language) — it acts as a medium between them.
Algorithms, which are written in a computer-understandable language (that can be any programming language such as Python, Java, or C++.….) acts as a medium between us and machine in the case of MACHINE LEARNING.
Father questioning Daughter again?
After some time, Father and Daughter again stood on the balcony and their conversation followed in this way:
Father: My dear!! “What is that object that is parked there called ?”
Daughter: I don’t know dad!! This is looking different, which I haven’t seen before.
Father: Ok, try this one, My dear!! “What is that object that is parked there called ?”
Daughter: I also don’t know this dad!! This is also looking different to me, which I haven’t seen before.
These both cars are different from the one s’, which she learned previously from her dad (one car is having only three wheels and another car is very long)
She has failed to answer these questions because she is not trained enough (hasn’t learned enough about different types of car objects) as she just started learning.
She learns slowly as she grows and years pass by, then she will be able to recognize based on her learning experience.
The same happens with machine
If we don’t train(teach) the algorithm, then it will also fail to identify/recognize the above objects as cars. If we make machines learn with enough examples of cars (which is nothing but experience in the case of girls).
If the algorithm is also trained well enough, then it will also try to produce accurate results based on its training experience.
In simple words, we can say that…..
Making a machine learn things, which is instead done by humans is called ‘Machine Learning’.
This definition is completely subjective and this can only be considered for basic understanding, it can be modified even more and more, once you dig deep into the subject.
My understanding of Machine Learning would be
Instead of humans doing the tasks, Automating the tasks intelligently with help of machines is called Machine Learning.
In this definition, there can again be an argument on — What is different between ‘Automating the tasks VS Automating the tasks intelligently’. This will be discussed in another blog.
