b'ArtificialARTIFICIAL INTELLIGENCE (AI)is the science of training machines to perform human tasks. The term was invented in the 1950s when scientists began explor-ing how computers could solve problems on their own. IntelligenceWe take for granted how our brains effortlessly calculate the world around us, every second of every day. AI is the concept that a computer can do the same. While AI is the broad science of immolating human learning, machine Delivers at aintelligence (MI) is a specific subset of AI that trains the machine how to learn from data. MI models look for pat-terns in data and try to draw conclusions much as you or I would. The machines are not being explicitly programmed by people, you can actually give them examples and they Global Scale will learn what to do from those examples.Googles Autonomous CarLook at Googles autonomous, self-driving car as a familiar example of AI applied to a routine task. It has lasers on the top which are telling it where it is in terms of the sur-rounding area. It has radar on the front, which is informing Experts believe that AI has the potentialthe car of the speed and motion of all the cars around it. to set the stage for the next GreenThen, several times a second, it uses all that data to figure out not only how to drive itself, but also to figure out and Revolution.Joe Funk jfunk@issuesink.com predict what potential drivers around the car are going to do. The car is processing almost a gigabyte of data every second. Algorithms, a sequence of instructions used to solve a problem, are developed by human programmers to instruct computers in new tasks. These algorithms organ-ize enormous amounts of data gathered by Googles cars sensors and stored in memory into information based on certain instructions and rules. Its an important concept to understand because in machine learning, it is the learning algorithmsnot computer programmersthat create the rules. Instead of programming the computer every step of the way, the MI approach gives a computer instructions that allow it to learn from data without new step-by-step instructions by the programmer. This means computers can be used for new, complicated tasks that could not be manually programmed. 24/ SEEDWORLD.COMDECEMBER 2019'