Human civilization since the stone age has been evolving with the adaptations that they are put in. They constantly learn from the people around them and also the conditions they analyze. In the machine era, what if a machine can learn to predict the outcomes by itself, without any kind of explicit interference and predict outputs consequently. This is where machine learning comes into the picture.
What fascinates me about machine learning is how it can be related to Human learning. Suppose, a child who is newly born does not know to speak but it adapts from the data of language that he is made to come in picture with it and it grasps the data and learns from it to speak. Machine Learning can be considered such intervened with human learning. ML is a subfield of AI (Artificial Intelligence).
Machine Learning in today’s world is the core of many new technological advancements, like the Tesla automobile’s autopilot system, Apple’s Siri, Sophia the robot etc.
Thus, machine learning can be defined as “The ability of a machine to learn itself based on the algorithm it is compiled and the trained model and provide decisions/prediction based on the computational results based on current data is said to be machine learning.”
Confusion between AI vs ML vs Deep learning
AI
The ability of a computer to think like a human being is said to be Artificial Intelligence. It is a broader concept of machine learning and more smarter way of carrying out tasks. You can think of a capable turing test to determine if the computer thinks similar to a human. If you are getting answer from siri while talking to it on your phone you are very close to it.
ML
Machine Learning is a subset to AI or current application of AI based on the idea that we should give the machine access to the data and let it learn by themselves. ML deals with extraction of patterns from the datasets. ML not only finds the rules for proper behavior but also change according to the changes in the world.
Deep Learning
Deep Learning is a subset of ML. It deals with the neural networks that are not performing well and training them with similar machine algorithms to get better result.
I hope this makes your difference in ML, AI and Deep learning clear.
How does Machine Learning work?
Now, Lets see how Machine Learning works?
One of the approach can be where the machine learning algorithm is trained using unlabeled or even a labeled training dataset to produce a model. Then further, a new data is introduced to this machine learning algorithm which further lets it make a prediction based on the new data. Then the accuracy of this model is tested and if it has an adequate accuracy it is said to be successful model and can be used further with new data. But, if the accuracy is not adequate then the ML algorithm needs to be modified and the model needs to be trained again.
This process is carried out several times until a feasible % of accuracy is obtained. Further the algorithm with the most acceptable accuracy is deployed. This is just a high level example, there are many other factors involved in real time accuracy determination.
To sum up, we went from what is machine learning to How it is different from AI and deep learning, clearing up various misconceptions along with How machine learning works
Thank you for reading till here. Please do contact me for further doubts. I hope you have understood.
TECHIE'S SCRIBBLING 