What is Machine Learning?
Machine Learning can be viewed as developing an application with minimum programming. Everyday apps that user use on their phones, Tablets and laptops developed in the following process:
- Someone come up with idea for the application.
- S/He call someone who can do programming (write his idea in the language that computer understand).
- S/He describes set of steps that application needs to work. For example, to check if our financial statements are correct; we take assets subtract liabilities and equity and we should have zero.
- We continue to explain more and more rules, and have the system that behaves in the predictable way.
With Machine Learning it is different! Without getting into any technical explanation here is my view on it for your review. Let’s say you want to teach 2 new employees how to check financial statements. You give one of them all or many financial statements that you have and tell them “figure it out, there is enough information here”. Then you go to the other employee and explain s/he everything.
We can clearly see how First employee will be inefficient and have harder time to understand financial statements and second employee will have better chance to makes total sense of financial statements. That is exactly the reason why this technology did not happen 10 years ago. Imagine the first employee is now a computer, which do not forget, can process millions of numbers and do not get tired. The only thing we are missing is a lot of data (thanks to internet we have a lot of it for basically everything) and cheap processing power (thanks to gaming industry and bitcoins for cheap and powerful GPUs). And as you noticed, all learning process is just happening on the computer without you or huge team of programmers doing it.
As the result, we can build bigger applications faster and cheaper. There are some issues with accuracy in this approach, but there is an issue with “bugs” with a traditional approach, so this part is remained the same.
Importance of machine learning and AI based projects.
AI and Machine Learning are supported by big players, such as:
- KPMG – published report “Rize of humans”, where company discuss how increased digital labor will affect job market in next 8 years. Picture bellow is great visual representation of the trend. As we can see, bubble on the bottom-left will be shrinking continuously as automation and robotization will continue to drop in price.
- Google (Alphabet) – claim to change its strategy from mobile first to AI first during “Google IO 2017”. Google is actually extremely good example, Sergey Brin (co-founder of Google) gave interview during world economic forum, where he said that google was not considering all technologies that are in core of modern techniques seriously fairly recently.
It is easy to imagine that all these technologies demand astronomical amount of investments, and only accessible to huge firms. My favorite example is how Japanese cucumber farmer created sorting mechanism for himself.
What is the point of Machine Learning For Business?
In our coming series of articles, we will go through actual projects within different industries. Our goal will be to explain results that you can have (what will be in your report, what is possible or not possible to do). We will tailor all material to be an easy read for everyone.
After seeing some examples, you will have better understanding of what is going in AI field. Coming up next is description of data research. When any data driven project is about to start engineer needs to go through the data, what is happening and why it might take up to 90% of time consumed on the project is coming up next week.
Please comment and let us know what you are interested in, what subject or industry do you want to visit in better depth. Share, comment or contact us in any way you like.