Original Source Here
Will Machine Learning be able to predict the future?
The First Question that Arrives is — What is Machine Learning?
I am pretty sure that you must have heard of Machine learning or Artificial Intelligence at some point in your life, be it in a movie, TV series or some science/tech channel. You must have seen robots or used chatbots, what the weather will be for the next few days and many more examples are there. All these have been possible because of Machine Learning.
Machine learning (ML) is the study of computer algorithms that improve automatically through experience. It’s a part of Artificial Intelligence. Machine learning algorithms build a model based on sample data, known as “training data”, to make predictions or decisions without being explicitly programmed to do so (Source Wikipedia).
How does Machine Learning work?
So there are three types of machine learning:
i. Supervised Learning: the machine learning algorithm is trained on labelled data.
ii. Unsupervised Learning: the machine learning algorithm can be trained on unlabelled data.
iii. Reinforcement learning: It features an algorithm that improves upon itself and learns from new situations using a trial-and-error method. Favourable outputs are encouraged or ‘reinforced’, and non-favourable outputs are discouraged or ‘punished’.
I will not go into depth because that’s not my motive.
So, before training any machine learning model, we need to identify whether it’s a Regression problem or a Classification problem.
Classification is the problem of predicting a discrete class label output. For example, one of the common examples of classification is detecting spam emails. A machine learning model can be trained with a set of spam-like emails labelled as spam and regular emails labelled as not spam.
Regression refers to the prediction of some quantity, it could be discrete or variable, for example, the price of a house based on its features.
While describing the regression problem, we came across a term called features. To understand what a feature is, let’s take a few examples:
1. Suppose we want to predict the weight of a person just by looking at them, what we will do is we will consider the height of the person or whether that person is stout or lean. These are the features on which the weight of a person is dependent.
2. Another example can be when you want to predict the price of a house. The feature, in this case, will be the location of the house, the area/size of the house, hospitality near it and many more.
Now when we come to the point of predicting the future, we come to the point of whether we can predict the future or not. Since we are generating 2.5 quintillion data bytes daily in 2020. While we use machine learning, we have a set of features and a target column. In our case, the target column will be what the person will do next or what choice he or she will make next, and the feature column will be what they have done so far like taking a bath, what they had for breakfast, and much more. In other words, what they have been doing all day. But here comes a problem, and that is how we are going to arrange the data or how we are going to collect the data. Collecting data will be possible with the merge of IoT and Machine Learning in the upcoming years, the data would be based on a daily life basis, not just being on the phone will provide us with data, that’s why we require IoT as well.
So let us make it a bit simple with a hypothetical situation. Suppose a person in a room has a bed, an almirah, and a window. The features can be when he wakes up, takes a bath, eats and other activities. After collecting the data and training it using a machine learning model, we can predict what he will do next.
Coming to the bigger picture, that’s where the main issue comes into play because now we have almost or near to infinite features and an infinitely large dataset on which we have to train our model. Can our machine learning model handle that? Maybe all these could be possible in the future since we don’t have strong AI and neither supercomputer which can examine/store an infinitely large dataset.
Trending AI/ML Article Identified & Digested via Granola by Ramsey Elbasheer; a Machine-Driven RSS Bot