Go back

US news roundup: 3-9 June

     

This week: a record-breaking supercomputer, trustworthy artificial intelligence, and Jet Propulsion Laboratory priorities

In depth: Plans to earmark a large chunk of National Science Foundation funding for United States regions that have historically received less investment have not been well received by universities, researchers and some high-profile politicians.

Full story: Uproar over plans to redistribute NSF funding
  



Also this week from Research Professional News

US science graduates enjoy ‘exceptionally high’ employment—Science, health and engineering PhDs outperforming general population, data show
 


 
Here is the rest of the US news this week… 

National lab supercomputer smashes record

The Oak Ridge National Laboratory claims to have built the world’s fastest supercomputer: a massive machine called Frontier. In a recent performance test carried out by the TOP500 organisation, Frontier demonstrated its ability to perform one quintillion calculations per second—a milestone that the US Department of Energy has been working toward for several years. As the first computer to officially break the exaflop ceiling, Frontier is probably the most powerful machine ever created. There are, however, reports that China already has two supercomputers operating at exascale, although these machines have not been tested as part of international efforts to rank supercomputing power. Plans for the construction of Frontier were first unveiled in 2019. The supercomputer is currently in its final testing phase and is on track to be fully operational early next year. 

Darpa seeks ideas for more-trustworthy AI

The Defense Advanced Research Projects Agency is asking researchers to submit proposals addressing how to build more robust, assured and trustworthy artificial intelligence systems. The agency’s latest AI programme, Assured Neuro Symbolic Learning and Reasoning, is looking to develop new neuro-symbolic AI techniques, which combine traditional rules-based AI approaches with modern deep-learning techniques. “Motivating new thinking and approaches in this space will help assure that autonomous systems will operate safely and perform as intended,” said Sandeep Neema, ANSR programme manager at Darpa. “This will be integral to trust, which is key to the Department of Defense’s successful adoption of autonomy.” The deadline for proposal submissions is 28 November

JPL chief seeks to tackle spiralling mission costs

The new director of Nasa’s Jet Propulsion Laboratory, Laurie Leshin, has revealed that one of the top priorities for her tenure will be keeping mission budgets under control. In an interview with news publication Science Insider, Leshin acknowledged that cost overruns were a problem for JPL. While the lab has carried out a lot of creditable science, its cost and schedule performance has not been so stellar, she said. “It can be hard to understand and write down early on all the requirements that drive cost,” said Leshin. “How do we get better at that and start to systematically look at where we are missing?” Leshin is the first woman to lead the JPL and began her role on 16 May. She was previously president of the Worcester Polytechnic Institute.