DAX+0,13 % EUR/USD-0,02 % Gold-0,08 % Öl (Brent)+0,86 %

Watchlist - gelistete KI/ AI Firmen weltweit (Seite 13)



Beitrag schreiben

Begriffe und/oder Benutzer

 

‘China’s Google’ releases its first AI chip


Baidu unveiled an artificial intelligence chip called Kunlun during its annual Baidu Create event on Tuesday. The company joins a raft of other Chinese firms in designing hardware tailored for machine learning.
Seven years: Kunlun is optimized for AI tasks including voice recognition, natural language processing, image recognition, and autonomous driving. The new design is 30 times faster than Baidu’s original FPGA-based processor from 2011, but the company says it’s not ready to begin mass producing it yet.
AI EVERYwhere: The company’s AI capabilities are developing fast. Its driverless minibuses will soon be deployed on power plant campuses in China​. EasyDL, a deep-learning tool, is enabling doctors to DIY diagnostic aids such as more efficient parasite screening.
The chip on China’s shoulder: Customized hardware will help Baidu scale up its AI capabilities at lower cost. This is, perhaps, the latest sign that the escalating trade tensions with the US are galvanizing China’s interest in fundamental technology.
https://venturebeat.com/2018/07/03/baidu-unveils-kunlun-ai-c…
http://create.baidu.com/
http://www.technologyreview.com/the-download/611382/zte-may-…
DeepMind’s AI agents are better than humans, @being your teammate

The Alphabet-owned company has developed a way for AI programs to team up effectively in a simple video game.
Team players: Researchers at DeepMind trained teams of AI agents to play a game of capture the flag in a modified version of the first-person shooter Quake III Arena. Teamwork is extremely difficult to develop effectively in AI programs, because it involves dealing with a complex and ever-changing situation. DeepMind’s agents were able to collaborate with human players, though—and the players say the programs are better teammates than most people.
Winning formula: The researchers used an algorithm that trains a host of agents in parallel using reinforcement learning, a machine-learning technique modeled on the way animals learn.
Game changing: It’s a promising advance, because AI programs will need to work nicely with each other. But the world in which these programs operate is incredibly simple. Demonstrating teamwork in the real world will be a lot more challenging.
—Will Knight
https://deepmind.com/blog/capture-the-flag/
https://en.wikipedia.org/wiki/Capture_the_flag
http://quake.wikia.com/wiki/Quake_III_Arena
http://www.technologyreview.com/s/603501/10-breakthrough-tec…
http://www.technologyreview.com/s/611536/a-team-of-ai-algori…
 Diskussion durchsuchen


Beitrag zu dieser Diskussion schreiben