Strategy games have long been a proving ground for artificial intelligence, but what could driverless cars learn from a hyper-violent video game?
John Breeden II is an award-winning journalist and reviewer with over 20 years of experience covering technology and government. He is currently the CEO of the Tech Writers Bureau, a group that creates technological thought leadership content for organizations of all sizes. Twitter: @LabGuys
Whether you elected to celebrate Memorial Day with somber reflection or summer barbecues and parties, I hope everyone had an enjoyable time without too much rain to dampen the mood. For this column, as we ease back into work over this mercifully short week, I wanted to keep it light and cover some interesting developments that happened regarding a few previous columns.
AlphaGo Retires as Champ
I became a huge fan of DeepMind’s AlphaGo artificial intelligence software when it beat a human master at the ancient Chinese game of Go last year. Go was chosen to show off advances in simulated human intellect because the 3,000-year-old game can only be won using a degree of intuition.
» Get the best federal technology news and ideas delivered right to your inbox. Sign up here.
The typical pattern matching computers are so good at, but which doesn’t really measure intelligence, won’t win games of Go, especially against humans who have mastered it. To win, AlphaGo had to show true human intelligence and a high degree of actual cognitive thinking.
Since its original victory, AlphaGo has gone on to win more matches, including beating another human master, Ke Jie, in a critical match over the weekend at the Future of Go Summit in Wuzchen, China. In fact, AlphaGo seems pretty much unbeatable, with lots of implications for how to apply its artificial intelligence to other fields, including critical ones like cybersecurity.
But Google, which bought AlphaGo developer DeepMind, announced the program was hanging up its tiles for good after the convention. Basically, AlphaGo has done all it can in the world of competitive Go playing.
DeepMind Co-founder and CEO Demis Hassabis hinted in his blog the technology behind AlphaGo could now be applied to other projects and fields, though he gave no specifics.
“We have always believed in the potential for AI to help society discover new knowledge and benefit from it, and AlphaGo has given us an early glimpse that this may indeed be possible,” he wrote after the retirement announcement.
NIST Helping with Presidential Cybersecurity Directive
The last time I wrote about the National Institute of Standards and Technology cyber framework, it was to ask each of you to consider contributing and commenting on the proposed 1.1 guidelines. Many of you did just that, with NIST receiving lots of great feedback on version 1.1 and some suggestions for future amendments.
NIST is still putting it all together for an official release, but in the meantime, President Donald Trump issued the Presidential Executive Order on Strengthening the Cybersecurity of Federal Networks and Critical Infrastructure. Having an updated framework would certainly help that effort, so NIST created the draft Interagency Report 8170, titled “The Cybersecurity Framework: Implementation Guidance for Federal Agencies.”
The draft document was written specifically to help agency heads respond to the executive order in a timely way while the new 1.1 guidelines are still being worked on. The draft document has a lot of good advice in it, which will probably not be modified too much before its official publication.
GTAV's Driverless Car Assistance
Over the long weekend, I got to play a little bit of one of my all-time favorite video games, “Grand Theft Auto V,” from Rockstar Games. GTAV, as it’s known to fans, is set in and around a realistic-looking simulation of a massive city called Los Santos, which approximates modern-day Los Angeles, plus the surrounding areas like Santa Monica and other towns up and down the California coast.
The game is hyperviolent, which has put it in the spotlight, with various people blaming it for contributing to the delinquency of minors, the downfall of American society, dogs and cats living together, etc. But it made the news recently for more positive reasons, namely because of a Bloomberg report showing how the game is being used by several organizations and companies to help speed the development of driverless vehicles.
When I last looked in on driverless vehicles, the big focus was adding the technology to trucks so they could drive down highways to deliver their goods without human assistance. Several states are working with companies on pilot programs, and New York recently announced it was moving ahead with actual testing on public roads as soon as possible.
But even in states like New York that are pushing ahead to field testing, it will still take a long time to develop driverless technology if that is the only opportunity to collect data. Simulations could instead allow scientists to run millions of trials constantly, and GTAV with its huge road network of coastal highways, crowded city streets and unpaved country roads can act as a perfect proving ground for the latest driverless programs.
Alain Kornhauser, a Princeton University professor of operations research and financial engineering who advises the Princeton Autonomous Vehicle Engineering team, told Bloomberg the highways and byways of Los Santos would never be a substitute for bona fide asphalt, but that GTAV “is the richest virtual environment that we could extract data from.”
It’s kind of ironic the game some lawmakers have tried to ban in the past is now helping various government and private efforts to develop driverless vehicles. If any of those organizations need help testing how simulated drivers react to, say, someone running around downtown Los Santos shooting rocket launchers, please let me know!