The DataLand
They say “Data is the new oil”
With an increase in the number of users in the virtual world, data has become an important tool and commodity in recent times.
Data is collected from everywhere and every second, from the moment you open your laptop or mobile phone till you switch it off data, is being collected, classified, processed to make your life easy.
For example, you may have noticed that if you choose to search online for shoes to buy a pair for an upcoming wedding the next site you open will give you ads on shoes.
These things can even manipulate you to buy stuff online.
Well, this is how things work nowadays!.
There are various techniques and algorithms to process this data and give a result.
So how it started?
Humans saw birds flying and wanted to invent something so that they could fly too. Many efforts were made, many inventions were invented, and eventually, airplanes came into existence that enabled us to fly from one place to another. The source of all motivation was from mother nature. We humans are so enthusiastic that we look at different things in nature and try to replicate them in our way. So some people thought that The human brain is also an amazing thing. It can identify objects, recognize patterns, classify things, and much more. What if a machine could do all this stuff? Wouldn’t that be cool? Today, we have come a long way in Artificial Intelligence. Many AI models are invented that could classify things, predict the future, play games better than humans, and even communicate with us.
How does it work?
To put in simple terms assume an artificial intelligence model as a child, you first train a child how to eat, how to drink you show him how to move your elbow and tell him how many times you have to chew after you bite.
To make him learn to distinguish colors you train him by showing colors on different objects or show them in a book.
This is exactly how a model works, you first teach the model how to process certain things and act on that thing which at the end gives you an output.
This processing part is done by computers which unlike humans only understand numbers that too only 1 and 0 even selfies which you took on a sunny day can be processed as a bunch of 0 and 1 fascinating right!
Let me give you two examples of basic understanding of machine learning which is a branch of artificial intelligence.
So the first example, Think that you have 3 sensors one measure temperature(thermometer) one measures pressure, and one the last senses sunlight.
You collect data on a minutes basis and store them into your excel sheet and now you want to you that would it rain based on the data collected in your excel sheet.
So these three things will become your feature to predict whether or not it would rain that day.
So you would now plot these three data on a graph and circle those data when it rains you would get a certain range of these three parameters where it would rain off course there would be outliers as rain also depends on various other factors, it won’t be perfect but you would get an idea of when it would rain or not. This plotting of graph and classifying data range to predict rain is one such example of machine learning.
You see how we can make our life easy with a bunch of sensors and a computer!
The next example is of lane-keeping assist feature in modern cars, for the folks who don’t know about this feature, what it does is it steers your car and keeps it in a single lane without you touching the steering wheel.
In this process what the car does is that it takes photos of the road ahead and identifies lanes present on the road and gives a response to the motor which is attached to power steering to either move the car left or right.
Pretty cool right!!
Automakers are now even trying to enhance these techniques to decrease the error to such a point that one day we would have a driverless car.
These were only two examples of artificial intelligence there are many other which we are currently using in our lives but just don’t notice, from separating spam mails from work emails or from having personalized ads pop-ups, from you Alexa to your weather prediction app all of these small things from which we get assistance on day to day basis use this new technology.
With the advancement of technology and some brilliant minds on earth, we have even tried to replicate a human brain.
A neural network is a collection of neurons/nodes interconnected with each other through synaptic connections. An artificial neural network looks something like this.
The inputs to the neural network are fed to the input layer(the nodes in light blue color). Each node in a neural network has some function associated with it, each connection/edge has some weight value. The inputs are propagated from the input layer to the hidden layer (nodes in blue color). Finally, the outputs are received at the output layer(nodes in black color).
Now, let’s briefly learn about a perceptron model, which is nothing but a single node of a neural network.
It was invented in 1957 by Frank Rosenblatt at the Cornell Aeronautical Laboratory, a perceptron is the simplest neural network possible: a computational model of a single neuron. A perceptron consists of one or more inputs, a processor, and a single output.
Maths ahead!!
This diagram summarizes the entire process of a perceptron learning algorithm. Initially, there are “n” inputs. These “n” inputs are multiplied by “n” random weights. Then this linear equation is added with a bias. Following which this result is passed through an activation function.
So, what exactly is an activation function and why do we need it? Well, most of the real-world problems involve multiple dimensions/factors and they can’t be represented by a simple linear equation and this is the reason this linear equation has to convert into a result in more than a single dimension. For this purpose, we pass the linear equation into the activation function. We have a variety of activation functions such as ReLU, TanH, Sigmoid, and SoftMax.
After we pass the equation through any of these activation functions, we would be done with the process of forwarding propagation, but the problem is we cannot be sure that the result we get is correct. Hence, we would have to compare this result with the observed value. The difference between the predicted value and the observed value is known as the error. The perceptron learning algorithm aims to minimize this error by updating the weights.
Nothing in this world is perfect, Well AI comes with its cons as well.
Concentration of Power
Further evolution of AI will most certainly mean a lot of power concentrated in hands of a few people or corporations.
Security Issues
The potential damage of cyber attacks will move to a new level. Imagine a hacker attack on an AI, programmed to perform a surgery or drive a school bus. Eek.
High Costs
Operation costs associated with building, maintaining, and repairing AI machines keep many industries away from replacing humans with robots.
Ethical/Moral Concerns
Although tech professionals work on ethical codes for machines, AI itself is unable to solve ethical dilemmas the same way humans do.
Artificial intelligence is the future and Depending on the intentions of those who control it, AI can either solve problems or create them.
Some brilliant cases make AI feel like a tech miracle meant to improve our lives in a myriad of ways. At the same time, we already have cases of AI-powered machines behaving like their creators never anticipated.
Artificial intelligence is controversial, and it is unpredictable.
Time will reveal the truth. Hopefully, someday artificial intelligence will become the greatest achievement of humanity, but not the last.
Ciao
PS- Feel free to give your suggestions.
