This week, self-driving Tesla had a fatal crash. Other than that – a lot about robots, can AI create an art, cloning animals and more!
More Than A Human
Ray Kurzweil and people like him believe the Singularity is just behind the corner and promise the new perfect world. They are very optimistic about the future. But sometimes you should listen to the other side to better understand the problem or vision. Like this video pointing out the problems with the idea of Singularity.
The artificial pancreas – a device which monitors blood glucose in patients with type 1 diabetes and then automatically adjusts levels of insulin entering the body – is likely to be available by 2018. Currently available technology allows insulin pumps to deliver insulin to people with diabetes after taking a reading or readings from glucose meters, but these two components are separate. It is the joining together of both parts into a ‘closed loop’ that makes an artificial pancreas.
Mike Loukides and Ben Lorica examine factors that have made AI a hot topic in recent years, today’s successful AI systems, and where AI may be headed. This article requires you to have an O’Reilly account, which is free.
DeepMind got a job! They will use the deep learning algorithms to fight blindness. The company will get an access to 1 million anonymised eye scans to train a neural network to identify early signs of degenerative eye conditions.
Our artificial intelligence systems are advancing at a remarkable rate, and though it will be some time before we have human-like synthetic intelligence, it makes sense to begin working on programming morality now. And researchers at Duke University are already well on their way. The team led by Vincent Conitzer is trying to find the patterns in our moral choices and translate this pattern into AI systems.
DARPA went full cyberpunk. They will hold a hacking content between seven autonomous bots running on supercomputers and visualize the process using graphics looking like a computer game or hacking sequence from a sci-fi movie.
Computerphille asks dr Mike Pound to explain the convolutional neural network. If you don’t know how they work I think this video might be a good place to fill the gap.
PBS Idea Channel fed a recurrent neural network with their previous episodes and then asked the system to generate a new one. You can give it a try, but we warned, it’s a tough four minutes to digest. After 4:13 Mike explains how the episode was created.
And here’s another video by Idea Channel. The computers are able to create art, but the question Mike asks and tries to answer is are we going to allow the machines to do art? Definitely worth watching.
And the last thing on art and machines doing art. Blaise Agüera y Arcas, principal scientist at Google, starts with how neuroscience influenced early AI researchers, explains how a neural network works and how by reversing the process his team ended up with that famous psychedelic computer generated art.
As harsh as it might sound, it was just a matter of time when someone gets killed in an accident when the car was in a self-driving mode. I hope the lessons will be learned from this accident and it will not stop research on self-driving cars.
If you live in London and you order a takeaway, it might be delivered by a robot. Trial delivery from a limited selection of restaurants should start this year.
Google cars are able now to understand cyclist’s hand signals as an indication of an intention to make a turn or shift over. I wonder if it knows the meaning of other gestures a cyclist can use.
Here’s a video of MARLO, a bipedal robot from University of Michigan, taking its first steps on a wave field.
It may sound terrifying, but it isn’t what you might have in the mind (Skynet). Scientists at the University of Zurich are teaching robots to “hunt” so that the robots could potentially take a look at their environments and then discern a target in real time. They imagine to use the software to create a future luggage or shopping carts that follow you (I’d call such luggage Chester).
When it comes to relations between humans and robots, there are three main areas that spark most controversy. First one is robots taking jobs. Second is military autonomous killer robots. The third one are sex robots. Why it’s so controversial and what the consciousness has to do with it is explained by this article.
In his famous Robot series of stories and novels, Isaac Asimov created the fictional Laws of Robotics. Although the laws are fictional, they have become extremely influential among roboticists trying to program robots to act ethically in the human world. Now, Google has come along with its own set of, if not laws, then guidelines on how robots should act.
Uber decided to hire Knightscope K5, Dalek-like robots, to patrol its parking lot, because they were cheaper that human guards. It will be wheeling around the lot, on the look-out for trouble and, hopefully, it won’t use the Exterminate! program.
Warning! Serious dive into uncanny valley ahead!
When Amazon bought Kiva in 2012, Jeff Bezos decided to use the robots for Amazon and Amazon alone, ending the sale of Kiva’s products to warehouse operators and retailers that had come to rely on them. Fours years later, the gap in the market is starting to be filled out by new startups offering warehouse robots.
BAE, the British aerospace defense giant, recently had come up with some interesting/crazy ideas. The newest one is to instead of designing a drone in a traditional way (which may take months to complete) grow them in a process BAE calls “chemputing”, where the drone is grown in a vat, with the electronics made and embedded in the body.
A motor-assisted glove designed to let astronauts carry out mechanical repairs without suffering muscle fatigue is making its way to Earth. The glove uses actuators and artificial tendons to mimic the muscles of the human hand to augment the wearer’s strength and endurance.
There is a lab in South Korea that has cloned over 800 dogs since 2006, offering the service to bring your dead dog back for $100,000. Apart from their popular dog cloning service, they also clone cattle and pigs for medical research and breed preservation.
20 years ago scientists from Scotland successfully cloned the first mammal – a sheep named Dolly. This article from Scientific American checks how much have we done in the last 20 years and where we are now with the cloning technology.
A startup from New York has a bold plan to grow leather in a lab. The process they use avoids waste, cruelty and dangerous carbon pollution. And, because lab-grown leather comes with none of the hair or fat found on animal hide, producers use fewer of the toxic chemicals needed to treat natural leather.
Researchers from Norway created a tiny maze and then they put there single- celled euglena and ciliates (taking the role of Pac-Man) and larger, multicellular rotifers (the carnivorous ghosts) to play one of the most bizarre version of Pac-Man.