Five different AI algorithms have teamed up to kick human butt in Dota 2, a popular strategy computer game.
Recommended for You
Researchers at OpenAI, a nonprofit based in California, developed the algorithmic A team, which they call the OpenAI Five. Each algorithm uses a neural network to learn not only how to play the game, but also how to cooperate with its AI teammates. It has started defeating amateur Dota 2 players in testing, OpenAI says.
This is an important and novel direction for AI, since algorithms typically operate independently. Approaches that help algorithms cooperate with each other could prove important for commercial uses of the technology. AI algorithms could, for instance, team up to outmaneuver opponents in online trading or ad bidding.
Collaborative algorithms might also cooperate with humans. OpenAI previously demonstrated an algorithm capable of competing against top humans at single-player Dota 2. The latest work builds on this using similar algorithms modified to value both individual and team success.
The algorithms do not communicate directly except through game play. “What we’ve seen implies that coordination and collaboration can emerge very naturally out of the incentives,” says Greg Brockman, one of the founders of OpenAI, which aims to develop artificial intelligence openly and in a way that benefits humanity. He adds that the team has tried substituting a human player for one of the algorithms and found this to work very well. “He described himself as feeling very well supported,” Brockman says.
Dota 2 is a complex strategy game in which teams of five players compete to control a structure within a sprawling landscape. Players have different strengths, weaknesses, and roles, and the game involves collecting items and planning attacks, as well as engaging in real-time combat. Pitting AI programs against computer games has become a familiar means of measuring progress.
DeepMind, a subsidiary of Alphabet, famously developed a program capable of learning to play the notoriously complex and subtle board game Go with superhuman skill. A related program then taught itself from scratch to master Go and then chess simply by playing against itself. The strategies required for Dota 2 are more defined than in chess or Go, but the game is still difficult to master.
It is also challenging for a machine because it isn’t always possible to see what your opponents are up to, and because teamwork is required. The OpenAI Five learn by playing against various versions of themselves. Over time, the programs developed strategies much like the ones humans use–figuring out ways to acquiring gold by “farming” it, for instance, as well as adopting a particular strategic role or “lane” within the game.
AI experts say the achievement is significant. “Dota 2 is an extremely complicated game, so even beating strong amateurs is truly impressive,” says Noam Brown, a researcher at Carnegie Mellon University in Pittsburgh. “In particular, dealing with hidden information in a game as large as Dota 2 is a major challenge.” Brown previously worked on an algorithm capable of playing poker, another imperfect-information game, with superhuman skill (see “Why poker is a big deal in AI“). If the OpenAI Five team can consistently beat humans, Brown says, that would be a major achievement in AI.
However, he notes that given enough time, humans might be able to figure out weaknesses in the AI team’s playing style.
Other games could also push AI further, Brown says. “The next major challenge would be games involving communication, like Diplomacy or Settlers of Catan, where balancing between cooperation and competition is vital to success.”
In early January, we predicted some of the biggest cyberthreats the world would encounter in 2018. Almost halfway through the year, it seemed like a good time to revisit that forecast and see how it’s playing out.
Recommended for You
Here’s what we’ve gotten right so far …
One of my predictions was that we’d see more huge data breaches, and that hypothesis was proved pretty quickly. In March, exercise- and diet-tracking app MyFitnessPal said it had suffered one of the biggest breaches in history: hackers stole the usernames, e-mail addresses, and passwords associated with some 150 million accounts.
That made the breach even larger in terms of sheer numbers than the massive Equifax hack of 2017. The only silver lining was that many of the passwords were protected by strong encryption, which seems to have limited fallout from the attack. Then there’s the Facebook imbroglio with Cambridge Analytica, which blew up the same month.
Some 87 million users of the social network had their data shared without their knowledge or consent. Strictly speaking, this wasn’t a hack. But I think it merits a (dis)honorable mention here because had the social network put tighter controls in place, it could have spotted the unauthorized use of the data faster and stopped it.
I also predicted even bolder efforts to steal computer processing power for cryptocurrency mining, and we dug into this risk in more detail later in January (see “Forget viruses or spyware–your biggest cyberthreat is greedy currency miners“). In the past few months, we’ve seen mining-minded hackers use popular malware such as Coinhive and Crypto Miner to hijack cloud computing capacity at companies like Tesla and British insurer Aviva. And one big security company, Darktrace, says it has found rogue mining software on the systems of a thousand of its customers.
Another forecast was that hackers would be likely to target more cryptocurrency exchanges. The latest assault happened earlier this month when Coinrail, a South Korean exchange, was compromised and almost a third of the coins it held were stolen. Maya Horowitz of security firm Check Point says it’s now seeing new cyberattacks on exchanges “every couple of weeks.” In response, security researchers and law enforcement agencies are stepping up efforts to track down the hackers (see “Sitting with the cyber-sleuths who track cryptocurrency criminals“).
… and kind of right
In January, I warned that ransomware attacks would cause even more damage.
These involve malware that locks down computer files with strong encryption and decrypts them only after a ransom has been paid, typically in untraceable cryptocurrency. I thought ransomware would cause a headache for big cloud providers like Amazon and Google, but the big story so far in 2018 has been the huge attack on the US city of Atlanta, which paralyzed a wide range of its municipal systems. The data kidnappers, who demanded £51,000 in Bitcoin, did some lasting damage, including erasing years of police video records.
Separately, I highlighted the potential risk of a significant cyberattack on physical infrastructure. I’m delighted to report that this prediction hasn’t proved correct so far, but the US Department of Homeland Security, the FBI, and Britain’s National Cyber Security Center did take the unprecedented step in April of issuing a joint warning that Russian hackers are targeting routers and other network infrastructure at power grids and military installations.
Here are the wait-and-sees
To my knowledge, there hasn’t yet been any concrete evidence of hackers weaponizing artificial intelligence, which I forecast, but plenty of cybersecurity companies are on the lookout for it. And it’s too early to tell whether there’ll be a concerted effort to hack election infrastructure, particularly in America, parts of which are still vulnerable to cyberattack.
The real test will come during the US midterm elections later this year.
And here’s that embarrassing miss
No sooner had the digital ink dried on my forecast than news emerged of serious security flaws in some semiconductors made by companies like Intel and AMD. Dubbed Meltdown and Spectre, these affected billions of chipsand effectively made it possible for hackers who’d already compromised computers to get access to secure portions of processors, where they could install malware or steal encryption keys. There’s since been a massive, and ongoing, effort to address the problem through software fixes and planned hardware changes, though new variants of the flaws keep popping up.
Apologies for not seeing this scenario in our crystal ball.
It’s a humbling reminder that when it comes to cybersecurity, the risks don’t just lurk in code.
IP Security Camera,P2P network IP Security Camera,Built in Microphone IP Security Camera with Intelligent Motion Alerts,Support Remote View and Control via Smart Phone/Tablet/PC,2 Way Audio chatting,Colour Recording IP Security Camera for Home Shop Indoor Security white
Smart Home Security for your life. We believe anyone should be able to go anywhere, do anything, and still know what’s happening at home. That’s why we designed this camera. 1. Records video clips when there any motion events detected. You can also remotely view live video from Smart Phones, iPad 2. The Motion Sensor of Multi-zone and Adjustable sensitivity. If any motion detected, will Alerts through email or push notifications(iOS and Android) to keep you notified. 3.Utilizing advanced H.264 technology, uses minimal bandwidth and storage requirements with smooth real-time video and audio streaming. 4.wide angle viewing allows you monitoring your whole house with clear HD video. Fast and easy setup 1. Simply scan, connected, and See What you want! 2. View your camera online Within minutes. With the P2P Server Technology, no complicated port mapping or DDNS settings needed.It’s very simple and smoothly. Residential Surveillance Uses: (1)Monitor and secure interiors such as bedrooms,apartments,family rooms, nurseries,garages,guest houses,and more. (2) While away check in on pets, children and elderly persons. (3) Camera protects your home by deterring trespassers. Package Including: 1x IP camera 1x USB Cable 1x Manu for camera 1x Bracket
- Easy to Use: Quick WiFi setup via IOS or Android smart-phone/iPad using supplied APP.Remote access to the camera on your Smart-phone.
- Super Wide Viewing Angle: The sleek, attractive camera,makes it ideal for surveillance applications such as offices, retail stores, stockrooms, lobbies, homes, apartments and more.Always keep an eye on what matters the most, with an outstanding wide angle of view,enough to cover any normal sized room.
- 64GB Supper Large local storage: you can share the exciting moment to your friends and family with the Large storage.
- Motion Detection:Get informed about every movement in your house,The motion detection sensors ‘feel’ it when something is moving in the room and will let you know immediately. It will take a picture and push a message on your phone. Perfect for surveillance!And built-in auto IR-CUT filter can provide clear and exquisite image with true display.
- Our supplies Best service attitude for you.any problem,please feel free to contact us!
Discounted: Sale Category