This week – Disney’s acrobatic robots; DeepMind’s AI learned teamwork by playing Quake III Arena; teaching robots to hug; augmenting human intelligence; and more!
More than a human
Here is a series of short lectures presenting how we can use AI not to replace us but to improve ourselves. The first one (starts at 7:06) discusses how combining human and machine intelligence gives better results than just using human or machine. The second one (starts at 26:53) looks at merging mind and technology from a philosopher’s perspective and how it impacts our cognition. The last one (starts at 51:04) looks at how we can improve our performance and attention with technology.
Artificial general intelligence (AGI). Some say this will be humanity’s last invention. Some believe it will bring our doom. Other can’t wait for it and the promise of abundance. This rather long article (there is a 40 minutes long audio version of it) takes the recent achievement of AI research and mixes it with visions of the future from movies and books to paint possible outcomes of inventing AGI.
Google’s DeepMind shared the results of research and experiments in which multiple AI systems were trained to play Capture the Flag on Quake III Arena. An AI trained in the process is now better than most human players in the game, regardless of whether it’s playing with a human or machine teammate. The AI, named For the Win (FTW), played nearly 450,000 games of Quake III Arena to gain its dominance over human players and establish its understanding of how to effectively work with other machines and humans.
Should general AI have human rights? This online discussion gives you some pros and cons for you to make your decision.
Engineers from Disney are working some very interesting robots. In a short video (which you can watch here), they presented some of their acrobatic robots. Some of them look like sticks with joints. Others are humanoid robots that do stunts in the air. Pretty impressive.
JD.com, one of China’s big e-commerce companies, has unveiled a warehouse in Shanghai that, it says, only keeps humans on board so they can service all the robots.
This article describes recent advances in creating robotic hands as dexterous as human hands. This is a complex problem to solve which combines not only the mechanics but also the learning process. Having a robot capable of gripping like a human would be pretty handy in many applications, like sorting or preparing orders (which is the objective of Amazon’s Robotics Challenge s)
In 2015, Gill Pratt, a former Pentagon robotics researcher, wrote that robot capabilities had crossed a key threshold. Improvements in electric energy storage and the exponential growth of computation power and data storage, he argued, had enabled robots to learn and make decisions informed by the experiences of other robots. Three years later and we see the demand for robots is increasing.
HuggieBot’s purpose was to find the perfect hug so that other robots what need to work closely with humans can hug you like a human. Inside the article, there is also an interview with the author of the research who explains more about why the research was done and what lessons the team has learned.