Machine Learning
This week I will do a series of posts on interesting application topics related to my startup. I will cover: machine learning (this post), computer vision, and applications in surveillance and medical image processing. These general areas are driving the innovation of many startups today in all sorts of categories. So hang tight and enjoy the coming week of application-driven posts.
At Georgia Tech, I did my PhD research in the area of computer vision. You can find my dissertation here: “Geodesic tractography segmentation for directional medical image analysis” and if you actually click that link you might be the 3rd person ever to do so after myself and my mother, lol!
Let’s start with a brief discussion of machine learning. For decades, researchers have refined the ability of computers to learn from acquired datasets and use that information to perform tasks. For instance, Gmail uses machine learning to figure out which emails should be sent to spam and which should not. It does a pretty good job because over the years as people have marked messages as spam Google’s computers have learned from those human decisions and now mimic human-like decision-making.
At last week’s GPU Technology Conference, my friend Bryan Catazaro did a live machine learning demo during the keynote. You can watch his presentation on YouTube.
Machine learning has many applications. I’ll cover surveillance later this week.
While the algorithms, principles, and math behind machine learning have been around for a long time, recent advances in the availability of rich data sources (Big Data) and heterogeneous computing have made these new applications tractable for startups to pursue.
What are your thoughts on interesting applications coming forward in the area of machine learning?