Hype around Machine Learning

In this article we are going to see what are the hypes and myths around Machine learning, How exactly the machine learning works and the limitations of machine learning and what next is superior to Machine learning.

Most of the people from other domains than Machine Learning thinking that Machine learning will take over the humans and everything will be automated, NO its not like that.

The difference between Machine learning and the traditional programming is that, We give the inputs and rules in the traditional programming(ex: 1+2, where 1 and 2 are inputs + is the rule), whereas in machine learning we give inputs and outputs and let it find the rules based on the relationship between inputs and outputs. The core concepts of all Machine Learning algorithms are derived from statistics and math.

To understand more about ML (Machine Learning), I will take the simplest ML algorithm which is Linear Regression and we will see how it works. The mathematical equation of Linear Regression is Y= mx + c, where Y is the value to predict based on the given x value. c is the intercept of the line and m is the slope of the line. In machine learning terms m and c are called parameters.

During our school days we would have studied the equation of straight line, The same concept applies here with some modifications. The linear regression algorithm first starts with the random line to fit the data points. After fitting the line it calculates the error between the fitted line and the data points by using the loss function such as Mean Squared Error(MSE).

The objective is to reduce the error so that the parameter m and c will be best fitted to the data points. To reduce the error we use a concept called Gradient Descent which finds the best parameters of m and c by calculating the derivatives of m and c with respect to error. This whole process goes number of times till the error reaches minimal.

Then we evaluate the trained model by using the mathematical functions like R2 and adjusted R2. To understand more mathematical concepts behind Linear Regression please follow this link.

From the above ML algorithm we can understand that what exactly happens in Machine Learning. so its not possible as of now to make a deadly Terminator which takes over the humans.

Machine learning algorithms perform well on numerical data, when it comes to processing of images and audio the ML algorithms struggle. To over come this issue Deep Learning is used. Deep Learning is the sub set of Artificial Intelligence. Artificial Neural Network (ANN) is the fundamental architecture of Deep Learning. Artificial neural networks (ANN) were inspired by information processing and distributed communication nodes in human brain. The Deep Learning is widely used on Computer Vision and Natural Language Processing fields.

Machine Learning is mostly used by the data scientist to find the solutions to the various business problems. The biggest cons of machine learning and deep learning is that it requires large amount of computing power and high quantity of data.

So don’t worry Artificial Intelligence will not take over us unless Elon Musk comes up with the new idea :-). Inspired from memes to write this article.