Machine Learning

Machine Learning is a branch of Artificial Intelligence which focuses on the use of data to improve the accuracy of computer algorithms so that the machine itself understands how to relate the data together, find patterns and become better. This is much like what humans do by learning from real world events and experiences.

The working of machine learning can be broken down into 3 parts:

1) The Decision-Making Process

This involves the use of machine learning algorithms to make a prediction for a given input data that is provided by the user. This prediction’s accuracy may vary depending on the amount of data and the experience that the algorithm has had with such data beforehand.

2) The Error Function

The error function evaluates the prediction made by the algorithm and if previous data of that type has been given to the algorithm, the function can draw parallels between the current data and all the previous data to assess the accuracy of the result provided by the algorithm.

3) The Model Optimization Process

So, the algorithm makes a model as it gets different sets of data and it keeps tweaking the model every time to increase the accuracy of the predictions given by the algorithm until the model becomes nearly perfect and gives accurate predictions consistently.

Machine learning is of two types:

i) Supervised Machine Learning

In this type of machine learning, a machine is helped to devise its initial model by using data that the user has already used and knows the outcomes of. This is use to train the machine faster as compared to using unknown data whose results are not even known to the user. An example of this is teaching an algorithm to filter out malicious and just random spam emails from your inbox and send them to the spam/trash folder.

This is similar to how schools always give you a few examples of how to use a mathematical formula before telling you to implement it in unsolved problems.

ii) Unsupervised Machine Learning

In this type of machine learning, the machine is given a large amount of unlabeled or unsorted data and the algorithm’s job is to make an initial model all by itself and find links or connections between various data sets. This is very useful in tasks such as data analysis, and image and pattern recognition.

This is similar to how you are given practical problems in science classes and asked to find the most appropriate solution without taking any external aids. It basically gives you no limits and you can use any method you see fit but there will always be a way for you to perfect your method further and improve it.

Machine learning has vast implications in real world. Listed below are some of them.

- AI chatbots configured for taking customer feedback or help customers on online websites.

- Speech-to-text converter as well as speech recognition software uses machine learning to analyze our speech. Voice search assistants like Alexa, Google Home, etc., also use such algorithms.

- Recommendations in your social media feed are also a result of machine learning algorithms which analyze your recent searches and show you content related to that to maximize your engagement time on the website.