The exhibit below gives detail of Uber’s patents on machine learning. 9/61 patents are on Ensemble Learning. It is a machine learning process that helps obtain better predictive performance by training multiple algorithms. 7/61 patents are on the Kernel Method – a class of algorithms for pattern analysis – and the rest is on Machine learning models.
The numbers in the exhibit are calculated based on CPC analysis. It represents the number of patent families classified in a CPC. A patent family is a group of patent documents –globally or in a jurisdiction – that covers the same (or similar) information.
In the patent portfolio of Uber, there were 2496 patent documents that belong to 884 patent families. The 884 patent families were categorized into 1351 CPCs. More often, a single patent gets categorized into more than 10 CPCs.
