Foundations of Machine Learning


Artificial Intelligence

Statistical Machine Learning

© 2017 - Mike Ashcroft

Fundamentals of Machine


Delivered through Persontyle

The following comes a post on the course here. For more information check out this very nice brochure.

The course was first delivered in November 2014 and was a lot of fun, and the response was very positive (see the evaluation results). I'll keep you posted on dates for future runs.

Despite the hype and confusion surrounding Data Science, the need for people who can interpret data and use it to find patterns and predictions to help organizations make informed business decisions is very real. Data Science is fueling the digital economy, we need to move it to the very center of our business, research and social change endeavours. Data Science is bringing new levels of speed, relevance, and precision to the way we design and manage businesses and operating models. Machine Learning is without a doubt the core aspect of Data Science and predicative analytics in general. In health care, Machine Learning is changing the way doctors identify people at risk of developing certain diseases; in retail, machine learning is used to analyze purchasing data to anticipate trends; CRM and marketing experts use it to tailor campaigns and offers.

Machine Learning is simple: We have the algorithms, we have the experience and, these days, we have the data. The ‘complicated’ mathematics behind the data revolution is the process of the cumulative application of basic techniques that can be understood in terms of mathematics we learnt in high school and early college. Providing this understanding is the most important facet of Data Science education: Only those who understand the tools they use are able to choose the appropriate technique for the tasks they face. I have never met an organization prepared to trust their data analysis to analysts who cannot explain why they use the techniques they do.

Fundamentals of Machine Learning bootcamp is designed to give you this understanding. It will provide you with the ability to apply the most powerful techniques in Machine Learning, to select appropriate techniques for particular problems , and to say exactly what these techniques do and why they work in a way that is understandable to data analysis stakeholders. MLB_Persontyle

Fundamentals of Machine Learning bootcamp will take you through the conceptual and applied foundations of the subject. Topics covered will include Machine Learning theory, types of learning, techniques, models and methods. Labs are developed to practically learn how to use the R programming language and packages for applying the main concepts and techniques of Machine Learning. In this bootcamp, our goal is to give you the basic skills that you need to understand Machine Learning algorithms and models, and interpret their output, which is important for solving a range of data science problems. This is an applied Machine Learning course, and we focus on the intuitions and practical know-how needed to get Machine Learning algorithms to work in practice, rather than the mathematical equations and derivatives.

Using actual data, the bootcamp begins by reviewing important basic statistical methods. You will learn to use the popular statistical programming language R to build these simple models from the ground up. You will then see how these simple techniques can be improved, combined, augmented and adjusted to produce powerful statistical tools for different tasks in data analysis. In this way, you will learn to see advanced Machine Learning techniques not as black boxes, but as principled techniques used to unlock patterns from data.

Over the course of five days, over two dozen techniques will be examined, implemented through supervised exercises and tutorials, and compared. You will learn the relative advantages and disadvantages of different types of techniques in different contexts. You will see how some models are entirely data driven, while others can be used to encode defeasible expert knowledge. You will learn methods for validating selected models and techniques and for choosing among alternative methods.

As we proceed we discuss with examples the sorts of data that suit these different approaches, and you will continue to apply these techniques ‘live’ in R. All topic areas have practical exercises, where you implement the algorithms we are looking at, as well as analyze their outputs and their suitability to particular problems. It is an essential part of the course aims that you get real ‘hands on’ experience working with the techniques we cover, in the comfortable environment of a classroom where you can discuss and work through problems you encounter with the instructor (me). The purpose is to arm you with a set of tools that you know how to apply, how to explain and when to use, as well as their theoretical background.