Many sports fans live for this time of year, with all eyes on college basketball. Much of the fun centers around the basketball brackets, but many go bust before the NCAA tournament barely begins. This year has been no exception. On March 16th, in a huge upset, #16 seed, University of Maryland, Baltimore County beat #1 seed Virginia. It's the first time in history that has ever happened. ESPN reported the UMBC stunning victory busted every single one of its 17 million brackets.
Could machine learning help people predict better brackets? Some experts believe it can. Machine learning is the process that makes predictions based on data. This means that anything in the realm of prediction—in basketball brackets and otherwise—would be a perfect fit.
A paper on this new method is published in the American Statistical Association Journal of Quantitative Analysis in Sports. The technique uses publically available basketball stats to look for potential upsets. Phys.org writes that these researchers used machine learning, decision trees, and causal inference in their method, which they use in predicting upsets in the NCAA basketball tournament.
We've seen machine learning and artificial intelligence reach industries like medicine and cybersecurity, and recently with Cisco's intent-based network. The company's network uses predictive analysis that can change the way an organization operates—spanning across data center, campus, branch and edge.
In a blog post, Cisco's Distinguished Engineer Debo Dutta writes the company is working towards improving everything in their core technology and implementing AI in security, collaboration tools, and IT. Working with Google's open-source machine learning framework TensorFlow allows Cisco to leverage public cloud in these efforts. In addition, Dutta writes that the company is also using Cisco's HyperFlex, a hyperconverged multicloud platform infrastructure.
Cisco's 2018 Annual Cybersecurity Report also states that more enterprises are using machine learning and artificial intelligence to track unusual behavior and patterns in web traffic. Machine learning algorithms can give a better view of user activity across multiple cloud platforms, helping defenders spot trouble before it happens. To learn more about this, check here.
We welcome the re-use, republication, and distribution of "The Network" content. Please credit us with the following information: Used with the permission of http://thenetwork.cisco.com/.