Watsonx, Generative AI attract US Open supporters

For more than three decades, IBM has aided the US Open in attracting and retaining fans as the event’s official digital innovation partner. The United States Tennis Association (USTA) and IBM Consulting collaborate every year to turn enormous volumes of data into insightful information for tennis enthusiasts.

The USTA is utilizing watsonx this year, IBM’s new business-oriented AI and analytics platform. With the help of a range of enterprise-grade, IBM-trained foundation models and classical machine learning and generative AI, watsonx enables the USTA to provide fan-pleasing AI-driven features considerably more fast. Users may work together on watsonx to specialize and deploy models for a range of use cases, or they can create their own models, enabling huge AI scalability.

AI-generated tennis commentary is powered by Watsonx

The generative AI capabilities of watsonx are being used by the US Open this year to provide text captions and voice commentary on video highlight reels of each men’s and women’s singles match. Each reel begins and ends with play-by-play narration, which is also audible during significant moments throughout. Both the US Open website and the US Open app will include the AI commentary function.

The first step in producing the commentary was to fill a data store on watsonx.data, a platform that links and controls reliable data from many sources (such as player rankings prior to the match, head-to-head records, match specifics, and statistics).

The teams then used watsonx.ai, a potent studio for training, verifying, optimizing, and deploying generative AI models for business, to train a foundation model. A wide range of contextual descriptions (such as adjectives like brilliant, dominant, or remarkable) based on rally durations, aces, first-serve percentages, relative rankings, and other important facts were incorporated into the US Open model, which was trained on the specific language of tennis.

Watsonx assists corporate clients in managing the complete AI lifecycle in addition to assisting them in integrating AI into their regular procedures. Because of this, the US Open will also employ watsonx.governance to organize, control, and oversee its AI-related operations. They will be able to operationalize and automate governance of their models, identify and reduce bias and drift, record and document model information, and encourage a collaborative atmosphere while also ensuring responsible, transparent, and understandable AI operations.

Watsonx is used to deliver a variety of Match Insights

Additionally, the US Open uses watsonx to give Match Insights, an entertaining collection of tennis data and forecasts, via the US Open app and website.

The IBM Power Index, for instance, combines performance and punditry to assess momentum. Using Watsonx.data and Watsonx.ai, structured historical data on each participant is merged with an analysis of unstructured data (language and sentiment gleaned from millions of news items about athletes competing in the event). As the game goes on, an additional 2.7 million data points are gathered from each shot in every game. This generates a rich, current data collection that can be used to perform predictive AI, forecast winners, and discover success factors. The Power Index offers insightful, relevant forecasts that provoke vigorous discussion.

When the tournament draw is made public, analysts and spectators frequently evaluate each player’s fortune and chances of moving up the leaderboard: did they have a “good draw” or “bad draw”? In order to assist them make more data-driven forecasts this year, IBM AI Draw Analysis provides a statistical component (a draw ranking) for each participant in the men’s and women’s singles competitions. Each player’s level of advantage or disadvantage is determined by the analysis, which is generated from structured and unstructured data using watsonx. The analysis is updated throughout the day as the tournament moves on and players are eliminated. Each player’s draw is graded from 1 (most advantageous) to 128 (most difficult), with 1 being the best. Additionally, fans may click on specific matchups to view the anticipated level of difficulty for that round.

Users of the US Open app may research a player’s path to the final and the difficulty of a player’s draw based on the AI Draw Analysis. Potential matches are displayed by the AI Draw Analysis tool using player information obtained from the Power Index. Fans may also follow IBM SlamTracker’s live stats and scores for ongoing matches.

Scalable corporate AI in a new age

Through their established cooperation, IBM and the USTA work together to investigate fresh applications for automation and AI to provide engaging fan experiences at the US Open. The developments from this year show how watsonx can assist businesses in swiftly and successfully implementing both predictive and generative AI technology. Watsonx opens the way to genuine AI scalability for the workplace by providing a collaborative, centrally managed environment that enables non-technical people to maximize the utilization of their organization’s high-quality data and utilize foundation models built on IBM-curated datasets.