What do you need to know about Machine Learning?
As we enter conference season once more, perhaps like me, you’re getting used to regular emails. Many contain advertorial content on Data Science, Internet of Things or more rarely Machine Learning.
Perhaps it is an indication that corporate life is more conservative, that Analytics and Big Data are still more prominent terms. I’ve yet to hear a Financial Services executive focus on Machine Learning as a key part of their insight strategy.
But, maybe it’s not just the world of business that’s removed from the interests of Data Scientists. Perhaps like the challenge science faces more broadly, it is considered too geeky/boring by many.
I was surprised to see how much, out of all these terms, Analytics still dominates Google searches (still far ahead of even Big Data):
That said, businesses are increasingly looking to hire Data Scientists. So, Data Science graduates are leaving universities, having been taught Machine Learning together with a mixture of Statistics & Computer Science. When I spoke with Data Science students at an event in Edinburgh, it was clear that they saw Machine Learning as a key part of their specialism, even if most businesses rarely mention it. That slow development is not a surprise to me, as it’s now 20 years ago since I was an R&D manager developing Artificial Intelligence pilots. In the two decades since, I’ve seen few businesses even attempt to apply the techniques I found so powerful (including Case Based Reasoning, NeuroFuzzy Logic & Genetic Algorithms). But, perhaps Data Science finally has enough momentum to take AI into the mainstream of commercial application.
So, if you’re looking to keep up with the developing Data Science or (wider) Customer Insight professions, what should you know about Machine Learning? Is it too late for you to learn? Do you need to return to university?
Although the social life options of the latter may sound appealing, most leaders do not have time to put their corporate careers on hold whilst they retrain. So, are there online resources to help you get up to speed & at least understand the language being used by your latest hires? The good news is yes and in this post I’ll share a few online resources & reviews that I hope you’ll find useful.
What better place to start than an online tutorial that claims to be the world’s easiest introduction. With the catchy headline of “Machine Learning is Fun!”, why not enjoy some of this two-part blog? Published on Medium by Adam Geitgey, it’s perhaps not as simple as some would wish, but it does provide a useful overview of techniques:
To balance that ‘data science perspective’ on machine learning, I thought it might also be interesting to share a market research perspective. This balanced & useful review by Kevin Gray in Quirks, provides just such a perspective. It should help researchers consider where AI algorithms could also be applicable to their quant work:
http://www.quirks.com/articles/2016/20160125-2.aspx
If all that education & advice has made you keen to get your hands dirty and try Machine Learning yourself, how can you get started? Well, if you are an R coder or have analysts in your team with R programming skills, here is a handy starting point shared by Jason Brownlee:
How To Use R For Machine Learning – Machine Learning Mastery
There are a ton of packages for R. Which ones are best to use for your machine learning project? In this post you will discover the exact R functions and packages recommended for each sub task in a machine learning journey. This is useful. Bookmark this page.
Don’t worry if you can’t or prefer not to use R, it seems that as well as a plethora of machine learning tools out there, there are some heuristics as well. In this quick start guide, Jason also shares how to understand any machine learning tool quickly (so good I included this second link from the same blog):
Understand Any Machine Learning Tool Quickly (even if you are a beginner) – Machine Learning Mastery
How can you learn about a machine learning tool quickly? Using the right tool can mean the difference between getting good predictions quickly and a project on which you cannot deliver. You need to evaluate machine learning tools before you use them.
Finally, to get you really ruminating over the subject, consider this more philosophical piece by Christopher Nguyen. In it he explores our relationship with AI the other way around. What can the ways machines learn teach us about our own brains, imaginations & the role of intuition. Thought provoking stuff:
http://adatao.com/blog/featured/2015/algorithms-of-the-mind/
So, I hope that was of interest. If you’ve discovered other great content online, to help us all better understand Machine Learning, please do share.
Have a great week learning more!