Your first encounter with AI from the comfort of your living room | NTT DATA

Wed, 26 June 2019

Your first encounter with AI from the comfort of your living room

In the last few years, we've seen a lot of progress in Artificial Intelligence that has given us a taste of the future and sparked many conversations about both the positive and negative impact of this technology.

AI is set to to transform every major industry and its impact will be comparable to the discovery of electricity or the invention of the internet. To quote Andrew Ng, one of the leading AI scientists: “AI is the new electricity”.

AI's high potential and impact requires a great number of experts in this field to respond to the needs of the labor market. 

This article intends to demystify AI and provide the necessary tools to understand it, without the need of spending too much money. To do this, we will start with some definitions and look into AI algorithms as well as  offer a list of resources to get you started on your journey to learn more about AI, from the comfort of your living room.


What are AI, Machine Learning and Deep Learning?

We're sure you already familiar with these terms, but let's start at the beginning. AI refers to the concept of building algorithms that emulate human behaviors and capabilities. The most widely used family of algorithms used for Artificial Intelligence is Machine Learning.


Simply put, Machine Learning algorithms use huge datasets to find patterns in the data, and uses these patterns to create rulesto to make predictions on new unseen data. The key advantage of Machine Learning algorithms is that you don't need to explicitly program any decision rules. The decision rules can be "learned" from the data. Machine Learning problems are usually divided into 3 categories:

  • Supervised learning: We provide a set of inputs and their corresponding outputs, and the machine learning algorithm tries to find rules to match the inputs with their outputs.
  • Unsupervised learning: We provide a dataset with inputs only and the algorithm tries to find patterns in the data. The most popular example of unsupervised learning is clustering.
  • Reinforcement learning: We train AI agents to take actions in an environment so as to maximize some notion of cumulative reward. These algorithms are used to train AI agents to play games such as chess and Go, and train robots to walk or execute specific tasks.

In the last decade, a new trend has emerged in Machine Learning called Deep Learning. Deep Learning refers to a class of artificial neural networks (ANNs) composed of many processing layers. ANNs have existed for many decades, but attempts at training deep architectures of ANNs failed until Geoffrey Hinton's breakthrough work in the mid-2000s. In addition to algorithmic innovations, the increase in computing capabilities using GPUs (general processor units) and the collection of larger datasets are all factors that have helped in the recent surge of deep learning. [5]

How do I get started with AI?

If you come from a non-technical background, the best way to start is to take the “AI For Everyone” course from You’ll need to dedicate around 10 hours to complete it. It was designed by Andrew Ng and is structured in 4 lessons:

⇒ What is AI

⇒ Building AI Projects

⇒ AI in Your Company

⇒ AI and Society

If you come from a technical background and you want a hands-on experience, then you’ll need some coding capabilities. Most of Machine Learning and Deep Learning courses use Python programming language.

If you don’t have experience in Python, then I recommend you start with one of the following resources:

⇒ Codecademy - Learn Python (Free with offers)

⇒ Book - A Whirlwind Tour of Python

  • Fast-paced introduction to essential features of the Python language, aimed at researchers and developers who are already familiar with programming in other languages.

⇒ Content as Jupyter notebooks (Free)

  • Before building a Machine Learning model, I recommend you become familiar with the Python data science stack (Jupyter notebooks, Pandas, Numpy, Matplotlib, etc…). The material below is a good introduction to the topic.

⇒ Wes McKinney - Python for Data Analysis Book

  • Jupyter notebooks (Free)
  • It includes a complete set of instructions to manipulate, process, clean, and crunchdatasets in Python. This hands-on guide is packed with practical case studies that show you how to solve a broad set of data analysis problems effectively. You’ll learn pandas, NumPy, IPython, and Jupyter in the process.

⇒ Dataschool - Best practices with pandas (10 videos) (Free)

  • Best practices with pandas to help students become more fluent at using pandas to answer data science questions and avoid data science errors.

Once you are familiar with Python stack for data analysis, you can start building your Machine Learning and Deep Learning projects. Below are some of the best resources for both:

Machine Learning

⇒ - Introduction to Machine Learning for Coders (Free)

  • Learn the most important machine learning models, including how to create them yourself from scratch, as well as key skills in data preparation, model validation, and building data products.

⇒ Andreas Mueller - Introduction to Machine Learning (Free)

⇒ Sebastian Raschka - Python Machine Learning, 2nd Edition

Deep Learning

⇒ - Practical Deep Learning for Coders, v3

  • Excellent course to learn Deep Learning.

As you have seen, there is a vast amount of information online that will help you get started with this technology. AI will be one of the areas that spurs most interest and will have most impact across sectors. It is a discipline that requires many technical and technological skills, but also human skills. As well as knowledge in mathematics and statistics, it’s important to learn programming languages and the possibilities the cloud offers. Finally, keep in mind that working in this field will always keep you outside of your comfort zone which means you will have to be patient and meticulous. 


How can we help you?

Get in touch