Some people think of it in awe, while some people think of it in shock. Some see it as a conspiracy theory, while some in total disbelief.
It reminds us of the Terminator – a machine sent back in time by machines to eliminate mankind. While this may sound a little convoluted and farfetched, turn your attention to Deep Blue, the iconic supercomputer which defeated the world chess champion Gary Kasparov in 1997. This was attributed more to brute force and good old-fashioned artificial intelligence (GOFAI) rather than the modern-day version of machine language.
In 2016, AlphaGo an AI program, also from IBM, beat Lee Sedol in a 5 match GO game (one of the toughest strategy board games), the first time a computer had defeated such a high ranked GO professional. AlphaGo used machine learning, and its successors AlphaGo Master, AlphaGo Zero and AlphaZero are far more powerful, each adding to the machine learning ability of its predecessor.
Machine Learning (ML) is basically an algorithm, a subset of AI, which provides the machine (or system) the ability to learn and improve from its own experience, rather than being explicitly programmed. In simple terms, it means that instead of writing the logic by coding, you just feed data to the algorithm to build the logic. Machine Learning allows us to gain insights from datasets; the machine learns to make sense of data. In businesses today, it is a powerful tool enabling you to make accurate and predictable decisions about customers, products, marketing efforts, recommendation engines and many other business applications.
A classical example of a machine learning based application is the wide array of digital personal assistants like Alexa, Siri and Google Assistant.
They have been created from a set of machine learning algorithms that updates the service over time based on user input, like new words and syntax. They all rely on natural language processing (NLP) to identify speech and synthesis, allowing them to recognize or speak words they have never encountered before. This is an iterative process and these digital assistants keep getting better over time.
Machine learning algorithms can be broken down broadly into three types:
Supervised Learning: Here humans act as the expert to train the machine to get the desired output.
Unsupervised Learning: There is not much human interaction as the machine learns by itself from the data fed into the algorithm.
Reinforcement Learning: The machine automatically gathers information from its interactions with the environment and tries to maximize the reward or minimize risk. Thus a feedback loop is incorporated into the algorithm.
Machine Learning – Is it the Right Career for You?
As we embark in 2019, AI, Machine Learning, Big Data, Deep Learning are not just buzzwords. They are being incorporated into our world at an amazing speed. They are fast becoming a way of our life, an integral part of our social fabric.
Uber, Amazon, Google, Netflix, Facebook, IBM Watson, HubSpot, Baidu, the list of companies using Machine Learning to facilitate their growth through more intelligent and personalized offerings is ever increasing. The importance of Machine Learning can also be gauged from the fact that as of December 2018, 5 of the 15 fastest growing jobs in the US involved Machine Learning and Big Data, according to LinkedIn’s 2018 US Emerging Jobs Report.
These are listed as below according to the growth from 2014 to today:
- Machine Learning Engineer (#2)
- Machine Learning Specialist (#4)
- Data Science Specialist (#7)
- Machine Learning Researcher (#14)
- Data Science Manager (#15)
Machine Learning: The New Path Ahead – The Best Books to Read
Machine Learning is here to stay, and is an excellent career move. This post lists a few of the best books one can read to get started.
Machine Learning covers the whole gamut of statistics, algorithms, programming, realistic data sets and case studies. Thus it is almost impossible to find a single book which covers it all. The list starts with a few books to get an introduction to Machine Language as a novice, followed by books catering to people who have some basic knowledge and would like to get advanced exposure to machine learning.
Most books focus on Python, by far the most popular and established programming language for machine learning. Some of the other languages used are R, Scala, Julia, SQL and SAS.
- Machine Learning For Absolute Beginners: A Plain English Introduction (Second Edition) – Oliver Theobald
This book provides a high-level introduction to the practical and statistical concepts of machine learning. As the title mentions, it is in simple-to-understand English, with no experience of coding required. The book includes clear explanations and visual examples. The second edition is not a sequel to the first edition, but a revamped and more up to date version.
- Machine Learning in Action – by Peter Harrington
Machine Learning in Action blends theory with practical real-life tools used in solving everyday problems. This is geared towards developers and a slight familiarity with Python would be useful. However, no prior experience with machine learning or statistics is required, as the book covers most of the concepts. A PDF copy of this book is available for download here.
- Programming Collective Intelligence: Building Smart Web 2.0 Applications – Toby Segaran
This is an easy-to-read and well-written book covering the fields of machine learning, data mining and optimization. Although it is a rather old book, it is a good practical guide for someone having their first brush with analytics. This book explains the concepts of collective intelligence using data and how to write code to analyze big data sets. You can download a PDF copy of this book here.
- Introduction to Machine Learning (Adaptive Computation and Machine Learning series) – Ethem Alpaydin
A good introductory book, it provides a broad overview of machine learning. The book covers the most important algorithms and techniques for a holistic approach encircling methods from computer science, neural networks, data theory and statistics. It has been used as a machine learning textbook at introductory courses in the undergraduate and early graduate levels. To download a PDF copy of this book, click here.
- Machine Learning: The Art and Science of Algorithms that Make Sense of Data – Peter Flach
As one of the most complete machine learning books available, this book provides a clear and example-driven approach to see machine learning in action. It introduces new and useful concepts with relevant background information. This is more of a contextual reading with a few statistics thrown in for good measure.
- Python Machine Learning – Sebastian Raschka
Python Machine Learning introduces you to the fundamentals of machine learning and how to use these in real-world applications using Python. The book classifies the various problem categories that machine learning can be used to solve. It provides background theory of various techniques followed the Python code, a good balance between theoretical math and coding implementation.
- Data Science from Scratch: First Principles with Python – Joel Grus
This book does a great job in walking through the tasks a data scientist would take to solve hypothetical problems, and explaining the models most popularly implemented in machine learning.
- Machine Learning: The New AI – Ethem Alpaydin
This book provides a concise overview of machine learning, its evolution, important algorithms and presenting real-life examples. The book also dives into areas of pattern recognition, applications of customer segmentation, reinforcement learning and the future direction of ML. Finally, the author talks about the future of ML and a discussion on the ethical and legal implications for data security and privacy.
- Fundamentals of Machine Learning for Predictive Data Analytics: Algorithms, Worked Examples, and Case Studies – John D. Kelleher, Brian Mac Namee, and Aoife D’Arcy
A very good book as an introduction to ML and analytics, this book describes the four approaches to machine learning: information-based learning, similarity-based learning, probability-based learning, and error-based learning. Each of these approaches is introduced by a nontechnical explanation of the underlying concept, followed by mathematical models and algorithms illustrated by detailed working examples.
- Pattern Recognition and Machine Learning – Christopher Bishop
One of the best books for beginners attempting to understand the Bayesian perspective on Machine Learning, it makes a good attempt to explain complicated theories in a simplified manner by giving examples and applications. A strong mathematical background would be helpful.
- Hands-On Machine Learning with Scikit-Learn and TensorFlow: Concepts, Tools, and Techniques to Build Intelligent Systems – Aurélien Géron
One of the most comprehensive books about machine learning, this book dives into two of the most important production ready Python frameworks, namely Scikit – Learn and TensorFlow. The language is very easy to understand and it is useful for both developers and non-developers. A newer edition of this book is scheduled to be released on 4th August 2019. A PDF Version of this book is available for download here.
Reading Your Way Ahead…
Reading books, the old fashioned way is a perfect way to complement any courses that you might take on machine learning. Instructor-led classes along with online tutorials help to hone the skills required for real-world implementation of machine language.