NEW YORK — When Roger Federer celebrates a high-pressure point over the next two weeks of the US Open at Arthur Ashe Stadium, IBM’s Watson cognitive computer will recognize that as a highlight in real time and alert the United States Tennis Association, which will then blast the highlight video out to fans through its social media feeds.
After testing the technology as a proof of concept at this year’s Masters Tournament and then putting it into action for the first time at Wimbledon, IBM has brought its Watson-powered “cognitive highlights” to the 2017 US Open.
Executives from IBM and the USTA say the automated technology gives them an ability to watch matches on seven courts simultaneously so that they can rapidly digest highlights and disseminate content to fans faster.
“This enables artificial intelligence, Watson, to automatically identify the best moments in a match and deliver that to fans,” said Kirsten Corio, USTA managing director of digital strategy. “You can get a sense of what’s happening point-by-point.”
ICYMI:
Sam Querrey went around the net post to hit a sweet winner past Sela during his straight sets win yesterday…#USOpen pic.twitter.com/Two6CsFHFo
— US Open Tennis (@usopen) August 31, 2017
Elizabeth O’Brien, IBM program director of sports and entertainment partnerships, said IBM has been teaching Watson to recognize player gestures and facial expressions, crowd noises and broadcaster reactions for the past two years across a spectrum of sports.
For the past nine or so months, though, Watson has been consuming video of tennis matches and been taught to read tennis highlights and lowlights in preparation for its US Open debut.
If Federer raises his fist to his face with a bent elbow, Watson will recognize that as a potential celebratory gesture, then cross-reference that with other unstructured data, such as the roar of the crowd, and structured data, such as statistical analysis, to determine the best moments of a match.
It will then auto-generate highlights, which helps to simplify the video production process and highlight-reel creation for the USTA.
“The US Open is packed with so much action across so many courts that even the fastest video team is challenged to keep pace with what’s happening,” said Noah Syken, IBM VP of Sports & Entertainment Partnerships.
In a sense, Watson is “watching” the game in the same way it can “read” X-ray and MRI scans, which have helped doctors discover ailments such as cancer and heart disease, said O’Brien.
“There are seven courts of play happening and every point is a potential highlight, so you’re looking at thousands of points,” she said. “With Watson, if gives you a manageable number.”
This year, the USTA is using Watson to create experiences that will drive fan engagement. A “Highlight of the Day,” as ranked by Watson, will be posted on its Facebook page, while video highlights will appear on player bio pages and at the fan-facing IBM Watson Experience on the US Open plaza. Fans that “favorite” players on the US Open app will also receive real-time push notification alerts about those players’ highlights.
Get The Latest Sports Tech News In Your Inbox!
IBM has even loftier ambitions for the technology over time, though. One day it may be used as a tool for broadcasters or help develop player skills during the training season, said O’Brien.
“We’re seeing this technology come to life through tennis, but the entire IBM Watson Media portfolio has the potential to impact many industries,” Syken said.
In January, the USTA opened a 64-acre, $63 million training facility in Orlando in hopes of developing future American-grown tennis champions. O’Brien said Watson might be able to help train athletes in Orlando by providing them with automated video from their training sessions that they can then dissect with coaches.
IBM is also behind SlamTracker, the official scoring app for the US Open that provides real-time scores, statistics, point-by-point analysis and data about player and ball position.
By being able to analyze things such as “depth of serve” or the distance a player might have traveled to score a point, and then cross-referencing that with historical statistics about things such as aces and unforced errors, SlamTracker might be able to complement the Watson video reels to help players at the USTA Orlando training center improve performance.