MODELING IN
PROBLEM-SOLVING
vs.
MODELING IN
ARTIFICIAL INTELLIGENCE
0
What’s the most popular adjective which comes to your mind right after saying
the word “model”?
SIMPLE
Why? Read the situation below.
Imagine there is an apple tree in your yard.
Due to the suitable gardening and direct reflection of the sun, the apples are fresh
enough. You want to pick an apple from the tree. There are 3 ladders in the basement.
To choose one, you need to measure the height of the tree. So, you pick a piece of
paper and start drawing the stuff you want.
There becomes a question
What will you draw
The sun?
The soil?
All of the flowers in your garden?
Or…will you only draw the parts which are necessary for the act of measuring?
So, you only focus on the tree, its branch (the one that the particular apple is
hanging from), and the apple.
Another question…
Should these 3 items be exactly as they are in the real world? Does it matter to the
drawing process?
Not at all!
Note that if there is an important feature for any of the items, it should be mentioned
already‼️
1
For the drawing process, you only need to:
Sieve the essential items.
Draw them as BASIC as you can.
This procedure which makes the process of problem-solving way easier is called “modeling”.
2
Now, forget about the previous situation and read the situation below.
Imagine you would like to implement a program that can choose the cat between the
pictures
What’s important to you through this process?
Specifying the details of the face of a normal cat as much as possible
Follow the next questions sequentially.
How should this program be implemented?
As a programmer, you are a mother and the program is your baby.
What does a mother do?
Training
What does a baby do?
Learning
What is the sex of this baby?
A program
This program is called a “model”.
‼️You are supposed to be THE MOTHER OF YOUR MODEL‼️
What is done to a model in the machine learning process?
A machine learning model is a program that recognizes patterns or behaviors based on
former experience or data. We train a model over a set of former data providing it with
a learning algorithm. The learning algorithm discovers patterns within the training data
and it outputs an ML model which captures these patterns and uses them to reason over,
learn from the former data, and finally make predictions on new data.
3
What do we mean by “the machine learning algorithms”?
The training tools which help you through this process are the machine learning algorithms.
4
Finally, look at the picture below in order to see the main parts of Machine Learning
process.