Welcome back to our course - AI for everyone
So far we’ve seen the framework of AI project cycle. Let’s see it with an example.
Imagine a scenario; a priceless diamond is kept safely in the famous museum for many years. One day a strange call come to the local policemen, in the call a strange person has threatened the policemen that someone is going to steal the diamond. The policemen had no idea who is that person in the call. No one is able to track that person.
Anyhow, they need to enhance the security of the diamond to make the area impossible to steal. For that they actually come out with an idea of using AI technology to safeguard the diamond from the strange person. Let’s see how it is possible.
As per the framework of AI project cycle, at first stage we have to write the problem statement that is, we have to scope the problem which we need to solve using AI project.
Here in this case the problem statement is, “Develop an AI enabled surveillance system to make sure that no unauthorized person enters the premises in which the diamond kept.”
Then as a second stage, we have to collect the data related to the project. In this case, we have to collect the photographs of all the authorized people who are all allowed to enter the premises.
Next in data exploration stage we need to represent the acquired data in a simplified format. For that we can create table and put the names and photographs of the authorized people that will give better and easy understanding of the data.
After exploring the data, we need to develop a system which detects the face of a person who enters into the premises and match it with the images of authorized people. If it doesn’t match then the AI system should alert the officials as an unknown person tries to enter the premises.
To implement this idea, in the modeling stage we have to look at different AI algorithms related to computer vision and select the best algorithm which will be suited for our project. Therefore at the end of this modeling stage our surveillance system will be ready.
Still we couldn’t deploy it immediately, before deploying the surveillance system, we need to test it with different known and unknown faces as input. By that we should ensure the accuracy of the surveillance system.
After proper testing, we can deploy the surveillance system in the museum. Now if anyone tries to steal the diamond then they will get caught in our system.
Yes!!!! Now we have successfully saved the diamond!!!! With the help our AI enabled surveillance system.
I hope you all understand the example.
Let’s recap
The situation is, a priceless diamond in the museum needs to be secured as the policemen got a threaten call to steal it.
Therefore, the policemen and the authorities of the museum trying their best to secure the diamond with the help of AI technology. For that we should follow 5 stages in general.
According to the situation in our example, in the first stage problem scoping we have to define the problem statement as, “Develop an AI enabled surveillance system to make sure that no unauthorized person enters the premises in which the diamond kept.”
In Data acquisition stage we have to collect photographs of all authorized people.
In Data exploration stage we have to store the collected data in the simplified format for better understanding.
Then in the modeling stage we have to select an AI algorithm to make the complete model of our project by implementing that algorithm.
At last in the evaluation stage we have to ensure the accuracy by testing our model and then deploy it in the real world.
Therefore, whenever we want to create an efficient AI project we should follow these 5 stages of AI project cycle. Okay!!
