There’s a lot of buzz about machine learning in government today, given its potential to improve operations, cut costs and produce better program outcomes. But what exactly is it?
Machine learning, or ML, is a collection of algorithms and mathematical models used by computer systems to progressively improve their performance on a specific task. Machine learning algorithms build a mathematical model of sample data, known as training data, in order to make predictions or decisions, without being explicitly programmed to perform the task.
Put more simply, machine learning is when computer systems use data to train themselves to perform certain actions – rather than requiring humans to do the same work.
Machine learning is part of a larger discipline called artificial intelligence (AI). However, it’s different than other types of AI in that it focuses on constant self-improvement – computers don’t have to be explicitly programmed but can change and improve their algorithms by themselves.
Machine learning is a big buzzword today, but it’s actually not a new concept. Artificial intelligence was first notably explored by Alan Turing during World War II, and the term “machine learning” was first coined in the 1950’s. Since then, programmers and statisticians have been seeking ways to improve machine learning’s performance and reliability.
The biggest hurdle to robust machine learning has been computing power. Until recently, programmers simply didn’t have enough computing power to fuel the robust data ingestion and synthesis required to make machine learning work.
You see, the accuracy of a machine learning model is based mainly on the quality and quantity of the historical data. With the right data, a model can analyze high dimensional problems with billions of examples, to find the optimal function that can predict an outcome with a given input. But it requires a lot of power to sift through, categorize and analyze that data.
It’s only recently – in the past 20 years – that such computing power has been available to organizations at scale. That’s largely due to the evolution of cloud computing which makes it easier and more cost-effective to leverage large computing environments.
The other barrier to using machine learning has been expertise. Over time, the algorithms behind machine learning use data and outcomes to improve. But those algorithms must first be developed – most often by software developers and programmers with highly specialized knowledge of statistics and artificial intelligence. Particularly in the public sector, those skill sets can be hard to come by.
But today, companies like Amazon Web Services (or AWS), are making it easier to deploy machine learning by taking their own algorithms and placing them within accessible, easy-to-use platforms and applications.
Amazon started AWS to allow other organizations to enjoy the same IT infrastructure, with agility and cost benefits, and now continues to democratize machine learning technologies to the hands of every developer. The structure of Amazon development teams, and the focus on ML to solve hard pragmatic business problems, drives AWS to develop simple-to-use and powerful machine learning tools and services.
These tools can put machine learning into the hands of every government organization, and with them, agencies can do amazing things. In our recent 10-minute course, How to Get Started with Machine Learning, we explain how those innovative undertakings can be achieved.