
Machine Learning Basics
The idea of making intelligent, sentient, and self-aware machines is not something that suddenly came into
existence in the last few years. In fact a lot of lore from Greek mythology talks about intelligent machines
and inventions having self-awareness and intelligence of their own. The origins and the evolution of the
computer have been really revolutionary over a period of several centuries, starting from the basic Abacus
and its descendant the slide rule in the 17th Century to the first general purpose computer designed by
Charles Babbage in the 1800s. In fact, once computers started evolving with the invention of the Analytical
Engine by Babbage and the first computer program, which was written by Ada Lovelace in 1842, people
started wondering and contemplating that could there be a time when computers or machines truly become
intelligent and start thinking for themselves. In fact, the renowned computer scientist, Alan Turing, was
highly influential in the development of theoretical computer science, algorithms, and formal language and
addressed concepts like artificial intelligence and Machine Learning as early as the 1950s. This brief insight
into the evolution of making machines learn is just to give you an idea of something that has been out there
since centuries but has recently started gaining a lot of attention and focus.
With faster computers, better processing, better computation power, and more storage, we have been
living in what I like to call, the “age of information” or the “age of data”. Day in and day out, we deal with
managing Big Data and building intelligent systems by using concepts and methodologies from Data
Science, Artificial Intelligence, Data Mining, and Machine Learning. Of course, most of you must have heard
many of the terms I just mentioned and come across sayings like “data is the new oil”. The main challenge
that businesses and organizations have embarked on in the last decade is to use approaches to try to make
sense of all the data that they have and use valuable information and insights from it in order to make better
decisions. Indeed with great advancements in technology, including availability of cheap and massive
computing, hardware (including GPUs) and storage, we have seen a thriving ecosystem built around
domains like Artificial Intelligence, Machine Learning, and most recently Deep Learning. Researchers,
developers, data scientists, and engineers are working continuously round the clock to research and build
tools, frameworks, algorithms, techniques, and methodologies to build intelligent models and systems that
can predict events, automate tasks, perform complex analyses, detect anomalies, self-heal failures, and even
understand and respond to human inputs.
This chapter follows a structured approach to cover various concepts, methodologies, and ideas
associated with Machine Learning. The core idea is to give you enough background on why we need
Machine Learning, the fundamental building blocks of Machine Learning, and what Machine Learning
offers us presently. This will enable you to learn about how best you can leverage Machine Learning to
get the maximum from your data. Since this is a book on practical Machine Learning, while we will be
focused on specific use cases, problems, and real-world case studies in subsequent chapters, it is extremely
important to understand formal definitions, concepts, and foundations with regard to learning algorithms,
data management, model building, evaluation, and deployment. Hence, we cover all these aspects,
including industry standards related to data mining and Machine Learning workflows, so that it gives you a
foundational framework that can be applied to approach and tackle any of the real-world problems we solve