|
|
Nvidia, University of Florida Plan to Build AI Supercomputer
|
|
|
|
|
|
Welcome back. Nvidia and the University of Florida are looking to develop the fastest artificial-intelligence supercomputer in academia. The move is part of a broader initiative between the chipmaker and the school to advance AI.
Editor’s Note: The WSJ Pro AI newsletter is making some changes. We will shift to a weekly publication schedule on July 28, and appear in your inbox every Tuesday going forward. The final daily edition will be published on Thursday, July 23. We believe that the newsletter, which launched in January, 2019, will benefit from a weekly perspective and a focus on even more analysis.
|
|
|
|
|
Artist rendering of what the new Nvidia-University of Florida supercomputer may look like. PHOTO: NVIDIA
|
|
|
Newsletter Exclusive: Nvidia, UF plan to build AI supercomputer. Nvidia Corp. and the University of Florida said Tuesday they aim to build the fastest artificial-intelligence supercomputer in higher education.
The supercomputer will be used to conduct research on rising sea levels, aging populations, personalized medicine, and other issues.
Graphics processors will provide the power. The new system will consist of 140 Nvidia DGX servers, which are designed for enterprise AI workloads. Each of the DGX servers has eight Nvidia A100 graphics processors. The supercomputer will have 700 petaflops of computing capacity. A petaflop allows for 1,000 trillion, or one quadrillion, operations per second.
UF to house AI tech center. The effort is part of a broader initiative between the chipmaker and the school. Nvidia also will collaborate with the university to develop an AI curriculum and coursework. And the university will house a Nvidia AI Technology Center, where UF graduate fellows and Nvidia employees will work to advance AI.
The two partners, said Jensen Huang, Nvidia’s chief executive, will “work together to make UF a national leader in AI and help solve not only the region’s, but the nation's challenges.”
The initiative will be funded by a $25 million gift from Chris Malachowsky, Nvidia’s co-founder and a University of Florida alumnus. Nvidia is providing $25 million in discounts on hardware, software, training and services. The university said it will invest $20 million in the effort.
|
|
|
|
|
The Collective and Augmented Intelligence Against Covid-19 alliance has developed an early version of a website aimed at sifting through the enormous amount of uneven information on the coronavirus.
PHOTO: EMAD MOSTAQUE/COLLECTIVE AND AUGMENTED INTELLIGENCE AGAINST COVID-19 ALLIANCE
|
|
|
Website aims to help global agencies make decisions on Covid-19. An alliance that includes an AI institute at Stanford has created a website designed to help global agencies such as the World Health Organization make better informed decisions on the coronavirus pandemic, The Wall Street Journal’s Sara Castellanos reports.
The website draws information from 80 Covid-19 sources, the World Bank and WHO’s Covid-19 database among them, said Emad Mostaque, the alliance’s chief architect. Information on the dozens of contact-tracing apps in use around the world is also included on the website.
Thus far the site, which is not yet available to the public, focuses on three topics. The first is the contact-tracing through mobile apps and AI of people with Covid-19. The goal of contact-tracing apps is to track, notify and isolate people who might have been exposed to the virus. The next subject is how to address misinformation on the pandemic, and lastly to identify marginalized areas most vulnerable to the pandemic’s health and economic toll.
|
|
|
|
|
Senate report says China developing ‘digital authoritarianism.’ A new U.S. Senate report accuses China of using its growing technology resources to develop “digital authoritarianism” to run surveillance and censor information around the globe, Reuters reports. New Jersey Sen. Bob Menendez, the top Democrat on the Senate Foreign Relations Committee, said the country is deploying technologies such as AI and biometrics to keep tabs on its citizens and manage information, according to the report.
|
|
|
|
MIT creates deepfake of Nixon reading Apollo 11 disaster speech. Artificial-intelligence experts at the Massachusetts Institute of Technology made a deepfake video of President Richard Nixon announcing that the Apollo 11 moon mission had failed, CNET reports. AI was used in the creation of Nixon’s facial movements, CNET says. The speech was read aloud by an actor. MIT's Center for Advanced Virtuality made the fake video to help people understand deepfakes and their impact, according to the report.
|
|
|
The U.S. government accused two hackers in China of targeting U.S. firms involved in coronavirus research, in a wide-ranging indictment that alleged the pair also stole hundreds of millions of dollars in sensitive information from companies around the world while working on behalf of Beijing’s premier spy service. (WSJ)
Facebook Inc. is creating new teams dedicated to studying and addressing potential racial bias on its core platform and Instagram unit, in a departure from the company’s prior reluctance to explore the way its products affect different minority groups. (WSJ)
|
|
|
|