In article 2, we introduced the various types of Machine Learning with the help of a bubble chart.
As you see from the above bubble chart, there are three main types of Machine Learning – Supervised, Unsupervised, and Reinforcement Learning. In article 2, we explained the two types of Supervised Learning – Regression and Classification. In this third and final part of the Machine Learning series, we will shed some light on Unsupervised and Reinforcement Learning.
Unsupervised Learning is the training of an AI algorithm using data that is not labeled, allowing the algorithm to act on that data without guidance. The algorithm finds natural groupings within the presented data, i.e. it identifies any hidden structure either by clustering or shrinking the data. There are two kinds of Unsupervised Learning – Clustering and Dimensionality Reduction.
Clustering is the process of using algorithms to identify how different types of data are related.
In Clustering, we’re just told, here is a data set. Can you find some structure in it? A clustering algorithm then finds natural groupings between objects within the given dataset. The algorithm groups the data into one or more data clusters based upon similarities, differences, or some other complex relationships between data objects.
In the example below, Clustering is being used for market segmentation. In the top row, a marketer is confused regarding whom to target to sell his expensive product. In the bottom row, Clustering finds similarities within the prospect base so that they can be grouped based upon their gender, income, age bracket, location, etc. Once grouped, the target market can be segmented. Customizing the marketing activity for specific segments delivers better business outcomes.
In addition to market segmentation, some other uses of Clustering can be seen in identifying fraudulent or criminal activities, identifying fake news, spam filters, recommendation systems, etc.
Dimensionality Reduction is the process of finding the natural groupings within the data by reducing the number of unnecessary features/variables within a dataset under consideration, thus obtaining a set of principal variables. Here the aim is to discard the features/variables that are not relevant to the target variable in any meaningful way and can be potentially considered noise. The Feature selection method is one way of Dimensionality Reduction. It is the process of identifying and selecting features that are relevant to your target variable. It can be done either manually by common knowledge or programmatically via various available tools.
Example: Consider that you are building a model that predicts people’s eyesight and you have a large amount of data that describes the person in detail – their biometrics, habits, demographic data, lifestyle information, details of their clothing preferences, etc. We can safely assume that the color of a person’s clothes or brand of shoes won’t be of much help in predicting the person’s eyesight. So, these fields can be dropped without hesitation. Consider lifestyle – do we think the number of hours they spend in front of a screen is important? Of course we do! That will be a factor in predicting their eyesight. By making simple manual feature selections, we have reduced the Dimensionality of the given data. This was possible because the unhelpful features were obvious or could be easily inferred based upon common knowledge. In case these features are not obvious, various tools can be used to aid the feature selection.
Of all the types of Machine Learning, Reinforcement Learning is closest to the kind of learning that humans and other animals do. As infants and toddlers, humans learn by interacting with their environment. This interaction produces a wealth of information about cause and effect, about the action and reaction, and about the steps involved in achieving goals, however small. This back and forth with the environment continues well into adulthood. Whether we are learning to drive a bicycle, or whether we are having a conversation with our colleagues, we are aware of how the environment responds to our actions and we seek to influence what happens by modulating our actions and behavior. Humans learn by trial and error. Reinforcement Learning is conceptually the same but is a computational approach to learning, by performing actions.
The typical framing of a Reinforcement Learning scenario: an agent takes actions in an environment, which is interpreted into a reward and an observation (representation of the state), which are fed back into the agent. Learning happens when an agent performs a sequence of actions based on decisions that will maximize the returned reward.
The goal of a Reinforcement Learning agent is to collect as many rewards as possible. In the most interesting cases, actions may affect not only the immediate reward but also the next reward, and through that, all subsequent rewards. The model needs to pay attention not only to the immediate reward but to the overall reward. These two characteristics — trial-and-error search and delayed rewards — are the two most important distinguishing features of Reinforcement Learning.
Reinforcement Learning is used in robot control, elevator scheduling, telecommunications, checkers, optimizing the driving behavior of self-driving cars, etc.
Over the last three blog posts, we have covered the basics of AI and Machine Learning. We started this blog series with a famous quote by Andrew Ng – “AI is the new electricity” – and just as electricity transformed everything as people knew then, AI is going to transform everything we know now.
AI and Machine Learning have already been subtly infused into the products and services that we all use regularly – email, Amazon shopping, Netflix, Spotify, Facebook, Alexa, Siri, robots, etc. Now that usage is further exploding to almost all industries, products and services.
Our goal was to provide a simple mental model – a framework for beginners, by which they could start understanding the basics of AI and Machine Learning. If you have gone through these three blogposts, you have indeed built a mental model, and further understanding of AI and Machine Learning will be fast. Carry on; learn from various leaders in the field. Don’t get left behind!