Until three weeks ago, Nvidia had planned to hold its GTC conference in San Jose, California, as usual, with CEO Jensen Huang delivering a keynote speech and various announcements. But the spread of COVID-19 forced Nvidia to discard those plans and switch to an online-only event.

Since then, hundreds of researchers, partners, customers, and Nvidia employees have been working remotely to produce GTC Digital, which kicks off this week, said Greg Estes, vice president of developer programs at Santa Clara, California-based Nvidia, in an interview with VentureBeat. GTC is founded on Nvidia’s graphics and CUDA processing designs, and it is intended for the “Einsteins and Da Vincis of our time.”

Roughly 30,000 developers have signed up. Normally, about 10,000 people show up for the live event.

Nvidia will still do hands-on training taught by seasoned instructors through its Deep Learning Institute. Those talks will now be done as real-time video sessions online from March 25 to April 10.

VB Transform 2020 Online - July 15-17. Join leading AI executives: Register for the free livestream.

“Starting Wednesday, we’ll make about 150 [precorded presentations] available, and over four weeks or so, that number will grow to 250,” Estes said. “We may be able to continue to expand the number of talks because once you’re in the digital format, we’re not constrained by the room capacity. So this is opening up some new avenues for us to look at.”

GTC typically packs hundreds of hours of talks, presentations, and conversations into a five-day event in San Jose. With GTC Digital, the idea is to take some of the best aspects of this event to a global audience and make it accessible for months. Some talks start on Tuesday, but most are happening on Wednesday and after.

Hundreds of speakers — among the most talented, experienced scientists and researchers in the world — have agreed to participate. Apart from the instructor-led, hands-on workshops and training sessions, which require a nominal fee, Nvidia is bringing most of the content to the internet for free.

“For those who had budgetary considerations, we’ve removed that as well,” Estes said. “If you’re a developer, it’s a big decision to come to a physical event. I’m seeing this as very much a silver lining to a difficult situation for everyone.”

The company refunded purchases for a GTC 2020 pass, and those tickets have been converted to GTC Digital passes. Passholders just need to log in with GTC 2020 credentials. Anyone else can attend with free registration.

Most GTC Digital content is for a technical audience of data scientists, researchers, and developers. But the company also offers high-level talks and podcasts on various topics, including women in data science, AI for business, and responsible AI.

The following activities will be virtual events that take place at a specific time (early registration recommended). Participants will be able to interact in real time with the presenters.

Training Sessions

Above: Digital art created for GTC Digital.

Image Credit: Nvidia
  • Seven full-day, instructor-led workshops, from March 25 to April 2, on data science, deep learning, CUDA, cybersecurity, AI for predictive maintenance, AI for anomaly detection, and autonomous vehicles. Each full-day workshop costs $79.
  • Fifteen two-hour training sessions running April 3-10 on various topics, including autonomous vehicles, CUDA, conversational AI, data science, deep learning inference, intelligent video analytics, medical imaging, recommendation systems, deep learning training at scale, and using containers for HPC. Each two-hour instructor-led session costs $39.
  • Live Webinars: Seventeen one-hour sessions, from March 24-April 8, on various topics, including data science, conversational AI, edge computing, deep learning, IVA, autonomous machines, and more. Live webinars will be converted to on-demand content and posted within 48 hours. Free.
  • Connect with Experts: Thirty-eight one-hour sessions, from March 25-April 10, where participants can chat one-on-one with Nvidia experts to get answers in a virtual classroom. Topics include conversational AI, recommender systems, deep learning training, and autonomous vehicle development. Free.

On-demand sessions

The following activities will be available on demand:

  • Recorded Talks: 150+ recorded presentations with experts from leading companies around the world, speaking on a variety of topics, such as computer vision, edge computing, conversational AI, data science, CUDA, graphics and ray tracing, medical imaging, virtualization, weather modeling, and more. Free.
  • Tech Demos: Demo videos narrated by experts, highlighting how Nvidia GPUs are accelerating creative workflows, enabling analysis of massive data sets, and helping advance research. Free.
  • AI Podcast: Several half-hour interviews with leaders across AI and accelerated computing will be posted over the next four weeks. Among them: Salesforce’s Kathie Baxter on responsible AI; Stanford professor Margot Gerritsen on women in data science and how data science intersects with AI; Ryan Coffee, of the SLAC National Accelerator Lab, on how deep learning is advancing physics research; and Richard Loft, of the National Center of Atmospheric Research, on how AI is helping scientists better model climate change. Free.
  • Posters: A virtual gallery of 140+ posters from researchers around the world showing how they are solving unique problems with GPUs. Registrants will be able to contact and share feedback with researchers. Free.

The tough thing is figuring out if conferences are going to be digital for a long period of time, Estes said.

“The first thing that we did was decide to take as many elements of GTC as we could and get them online,” Estes said. “We wanted to make them available to as many people as we could.”

There will be 150 posters that attendees and registrants can check out. Usually, the posters would be printed out and people would be able to browse them at the convention center. But they’re all being converted to be available online. Presenters are creating videos to show the tech demos and explain what they’re about. Altogether, Nvidia has about 1.8 million registered developers in its developer program.

“We’re very excited that 30,000 people have signed up,” Estes said.