Timeless AI insights at the GRAMMYs: A recap of the meetup experience
Looking more at the IBM Developer meetup hosted by the AI LA Community.
Nearly 100 software developers and artificial intelligence (AI) enthusiasts gathered on the evening of January, 28, 2020, for an IBM Developer meetup hosted by the AI LA community at Cross Campus in Los Angeles to learn about how AI enabled the GRAMMY webcast to take viewers deeper into the event than any previous broadcast.
On GRAMMY night on January, 26, the grammy.com telecast showcased Insights with IBM Watson, a new video-enriched experience that leveraged IBM AI and Cloud technologies to integrate insights into the GRAMMY Red Carpet livestream.
The meetup was a chance for IBM inventors and AI pioneers Aaron Baughman and Nick Wilkin to highlight some of the innovation that IBM applied to make this unique application work and enhance the GRAMMY experience for millions of fans.
During the evening, Matt Stark, Associate Partner in the IBM iX Business Services group, led Baughman and Wilkin through a discussion about how the app came to be and the technology used to build it.
IBM’s job was to enhance the fan experience on Music’s Biggest Night. IBM partnered with the producers of the GRAMMY awards, the Recording Academy, to contextualize artists’ experiences at the GRAMMY Awards with insightful facts. IBM technology, with the help of Watson Discovery, identified and shared the human journeys that led artists to a nomination at the GRAMMY Awards. Each of the insights needed to be timeless, unambiguous, interesting, and a magnification of a facet of the artist’s life. Now, fans around the world will be able to relate and understand the importance of their favorite celebrity’s accomplishment.
But for Baughman and Wilkin and the rest of the IBM team, the job was a deeply technical challenge.
A distributed AI pipeline was developed to support the creation of insights at scale with high availability. Prior to the event, AI was trained to:
- Gather over 18,000,000 pieces of content from 100,000 trusted sources such as news websites and blogs, Wikipedia, and grammy.com
- Perform both single-document and multi-document summarization
- Create factoids with natural language understanding and processing techniques
- Generate insights through deep neural networks built with IBM NeuNets and categorize them to be called on when needed
But the biggest challenge, as with any AI anywhere deployment like the GRAMMYs Insight with Watson, was to deliver computing at scale.
Watson processed more than 18 million documents into 350,000 factoids that were further refined into 1,068 insights for 900+ nominees. That effort required a flexible and computationally intensive platform.
Baughman and Wilkin walked attendees through the deployment that combined Docker containers and the Red Hat OpenShift on IBM Cloud platform to scale computational processing capabilities for batch processing. The entire AI pipeline was supported by 7 images running Node.js v12 and Python v3.7. The cluster itself had 6 workers, with 4 vCPUs and 16 GB RAM each. This configured cluster allowed IBM and the GRAMMYs to process all 900 nominees within 10 hours.
Wilkin also discussed the UX decisions that went into creating a functional admin tool that empowered grammy.com editors to make smart content decisions in real time during the broadcast.
Those who attended the January, 28, event viewed a presentation by the developers, in which they gave a demonstration of the early returns that Watson guidance offered in an enhanced webcast that took music fans deeper into the story of nominees than any previous broadcast. But the real significance was the promise of the tools and functions that Watson can enable in other business areas outside of the music industry, such as healthcare, financial, and sports applications.
Attendees enjoyed and engaged with the technology solution deployed at the GRAMMY Awards show and were immediately able to see connections to their own business and technology challenges.
For more detailed information on the solution, see Timeless AI Insights at The GRAMMYs: Changing the Fan Experience.