Artificial Intelligence is one of the emerging technologies that try to simulate human reasoning in AI systems. Researchers have made significant strides in weak AI systems, while they have only made a marginal mark in strong AI systems.
Most of us have used Siri, Google Assistant, Cortana, or even Bixby at some point in our lives. What are they? They are our digital personal assistants. They help us find useful information when we ask for it using our voice; we can say ‘Hey Siri, show me the closest fast-food restaurant’ or ‘Who is the 21st President of the United States?’, and the assistant will respond with the relevant information by either going through your phone or searching on the web.
Artificial Intelligence is the ability of a computer program to learn and think.
John McCarthy coined the term Artificial Intelligence in the year 1950.
He said, ‘Every aspect of learning or any other feature of intelligence can in principle be so precisely described that a machine can be made to simulate it. An attempt will be made to find how to make machines use language, form abstractions, and concepts, solve kinds of problems now reserved for humans, and improve themselves.’
How does Artificial Intelligence work?
Computers are good at following processes, i.e., sequences of steps to execute a task. If we give a computer steps to execute a task, it should easily be able to complete it. The steps are nothing but algorithms. An algorithm can be as simple as printing two numbers or as difficult as predicting who will win elections in the coming year!
So, how can we accomplish this?
Let’s take an example of predicting the weather forecast for 2020.
First of all, what we need is a lot of data! Let’s take the data from 2006 to 2019.
Now, we will divide this data in an 80:20 ratio. 80 percent of the data is going to be our labeled data, and the rest 20 percent will be our test data. Thus, we have the output for the entire 100 percent of the data that has been acquired from 2006 to 2019.
What happens once we collect the data? We will feed the labeled data, i.e., 80 percent of train data, into the machine. Here, the algorithm is learning from the data which has been fed into it.
Next, we need to test the algorithm. Here, we feed the test data, i.e., the remaining 20 percent of the data, to the machine. The machine gives us the output. Now, we cross verify the output given by the machine with the actual output of the data and check for its accuracy. While checking for accuracy if we are not satisfied with the model, we tweak the algorithm to give us the precise output or at least somewhere close to the actual output. Once we are satisfied with the model, we then feed the data to the model so that it can predict the weather forecast for the year 2020.
With more and more sets of data being fed into the system, the output becomes more and more precise.
Well, none of the algorithms can be 100 percent correct. None of the machines have been able to attain 100 percent efficiency as well. Hence, the output we receive from the machine is never 100 percent correct.