Wesley Morris, Vuzion Development Manager, writes …
When our marketing department asked if I wanted to write a blog I leapt at the chance. Which is unusual as I’m a software engineer by trade and it has been a long time since my grammar has undergone any kind of scrutiny.
But now I figured that I had a secret weapon at my disposal: Machine Learning. Artificial intelligence (AI) could compress thousands of hours of learning into a few minutes and make logical predictions based on what it learned.
So that’s exactly what I set out to do.
In the last few years, machine learning has gone from being a dark art of data science to a staple tool of the application developer’s toolbelt, thanks in part to services like the Azure Machine Learning studio.
Using some of my free Azure credit I spun up an Azure Function and imported in a Language Prediction model for my machine learning function to utilise. This model reads in a text prompt and then, using millions of articles as reference, tries to predict what the rest of the text should be.
So, with my solution up and running in the cloud, I fed in the following:
Machine learning will allow us to create smarter solutions to more complicated problems and forever change the way we view workplace automation.
And after taking some time to think, the AI replied:
The technology behind our smart machines is a natural extension of the human brain. Our brains are highly social creatures. We are drawn to and form ties with others, and the human brain is built with the ability to form connections with the Internet and the World Wide Web. What this means is that we have the ability to learn in real time and to make fast decisions with our minds.
This is the key advantage of a machine learning solution in the workplace. By combining the human brain with the knowledge base of the Internet we're able to leverage the power of this knowledge base to make decisions about how the machine should work.
The Internet is a huge and diverse collection of resources, with millions of topics available. This gives the machine learning solution the ability to search, learn, and learn again. The combination of the machine learning solution and the human brain is unique because we can leverage the power of our brains in real time to make intelligent decisions.
Looking at the text it’s apparent that while the AI can construct a natural flow of language, it can’t construct a narrative or reach a conclusion. Which means that my dreams of outsourcing to AI are still a few years away and this blog won’t actually write itself.
But the real takeaway here is how easy it was to utilise Azure to create quite a complicated AI solution and the breadth of options I found at my disposal. The Azure Machine Learning Studio has everything from guides for complete beginners to complex functionality for the most battle-tested data analyst. And Azure Functions with Logic Apps make deploying these AI services quick and production ready for surprisingly little cost.
If you want to explore what you can do with Azure, I highly recommend attending one of Vuzion’s Azure Fundamentals training courses, and our Azure Commercial Lead, Darren Roberts, will also be more than happy to help - call 0333 009 5939, or email email@example.com.