Just as it will impact so many other industries, the potential of Artificial Intelligence (AI) to transform sports is seemingly limitless.
The fields of analytics, venue management and content personalization are but three areas expected to benefit from added intelligence, but AI can also affect how sports are played at both a professional and amateur level.
Callaway Golf has used AI to develop its latest clubs, claiming the use of a Machine Learning algorithm and supercomputer helped reduce a tradition of 34 years of product design into a matter of just days.
AI-powered golf
“A couple of years ago a member of our R&D team started exploring the potential for AI to help design a better golf club," Dr. Alan Hocknell, Callaway’s head of R&D tells me.
"We decided to try the supercomputer and [an AI algorithm] out by designing the face of a driver, because the face of a club is instrumental to generating fast ball speed, which promotes distance, which every golfer wants more of."
Callaway designed the body of the club using traditional means and instructed the supercomputer to design a face that would suit the characteristics of the body. This means the face designed by the AI simply wouldn’t work as effectively if affixed to another club.
The ‘Flash Face' was the product of 15,000 cycles and has been included in a set of drivers and woods. Each cycle learned from the previous one, with the Flash Face boasting dozens of subtle internal ripples that increase ball speed and distance.
Hocknell says the creation of the Flash Face may not even have been possible without the use of AI.
A surprising outcome
“The process for improving a golf club’s design is no different than that of improving the design of almost any other product. Create a prototype, test and analyze it, tweak it, test and analyze it again etc.,” he continues.
“These prototypes can be physical or increasingly they are virtual prototypes which are tested in detailed digital simulations. But you’re confined to a human’s capacity for conception and analysis, which typically limits us to creating four or five iterations of a new part.
“Using AI and Machine Learning, our supercomputer worked for three weeks non-stop, creating 15,000 virtual prototypes, learning from each one, until it arrived at what it deemed the best design in terms of what we asked for, which was to deliver the fastest possible ball speed while at the same time conforming to the rules on face design maintained by the United States Golf Association (USGA), [one of] golf’s governing [bodies].”
What the AI eventually produced, surprised Callaway’s R&D team: “We were stunned at the final design, which is asymmetrical, with what appears to be a random pattern of waves and ripples of different sizes and depths spanning the surface. It bears no resemblance to any face Callaway has conceived before. By looking at it it’s hard to make sense of how it works.”
A worthy investment
There’s no doubting the willingness of golf fans to spend money on anything that will improve their game – The Epic Flash Drivers will set you back at least $530 – but has the investment in a supercomputer been worth it? Hocknell thinks so.
Our investment in the supercomputer was significant, but when you compare to humans the depth of its analytical capabilities, and the volume of work it did in three weeks, the bang for the buck is pretty good, he says. “According to our calculations it would take our R&D team, working with normal desktop computers, 34 years to create and analyze 15,000 prototypes – and even then it is unlikely we would have ended up at a design as radical as the one in Epic Flash.
“It’s possible AI can help us achieve other performance parameters we continuously seek to improve, such as increasing the stability and forgiveness of a clubhead, lowering backspin off metalwoods, increasing backspin off short-irons, and improving golf ball aerodynamics, to name a handful.”
