The Kynix Blog

 Inside a secretive AI nonprofit backed by Elon Musk and other Silicon Valley figures, a handful of robots designed to help out in warehouses are gradually learning how to do useful household chores.OpenAI, which was created to do basic AI research, is reprogramming robots developed by Fetch Robotics, a company that supplies warehouse automation hardware. Researchers at OpenAI are equipping the robots with software that lets them train themselves through trial and error.The effort reflects a bet that innovations in software and machine learning, rather than breakthroughs in hardware, are the way to give robotics remarkable new capabilities. Fetch makes a range of robots for warehouses, including systems that follow workers around a building, carrying items dropped into a basket. OpenAI is using a system that features a mobile base but also 3-D depth sensors, a 2-D laser scanner, and a robotic arm with seven degrees of freedom.In April, OpenAI recruited Pieter Abbeel, a professor at the University of California, Berkeley, and a leading expert on robot learning. Abbeel has shown how robots can use a machine-learning approach called deep reinforcement learning to acquire completely new skills that would be hard to program by hand, such as folding towels or retrieving items from a refrigerator. Google DeepMind, an AI subsidiary based in the U.K., uses this technique to get computers to play computer games at a superhuman level.Abbeel’s robots learn tasks from scratch, using a neural network that receives sensor input and controls physical movement. The network adjusts its parameters automatically as it inches closer to its goal. A robot might try thousands of grips, for instance, in the process of learning how to hold a certain object.“If this goal can be achieved, then there will be economic and industrial benefits,” says Marc Deisenroth, an expert on reinforcement learning at Imperial College London. “Imagine a Roomba not only cleaning your floor but also doing the dishes, ironing the shirts, cleaning the windows, preparing breakfast.”Deisenroth says using off-the-shelf robots could drive costs down. “Currently, the software seems to be the bottleneck,” he adds. “However, independent of this, better hardware could also lead to substantial improvements.” Soft manipulators and elastic feet similar to a monkey’s feet are concepts that researchers have started working on, he says.Some manufacturers, including the Japanese company Fanuc, are testing reinforcement learning as a way to train industrial robots quickly in new tasks such as learning to grasp unfamiliar objects. When many robots work in parallel, the training time required is reduced accordingly . Robot researchers at Google are testing similar learning techniques.“Moving away from having to program robots by hand by endowing robots to learn autonomously is a key element for the future of robotics,” says Jens Kober, an expert on robot learning at Delft University of Technology in the Netherlands. Kober says having robots share the information they have learned will be crucial.While robots such as those made by Fetch are finding their way into many factories and warehouses, domestic robot helpers remain the stuff of science fiction. Performing seemingly simple tasks like washing dishes or folding laundry in a messy home setting is incredibly hard for a machine. A robot programmed the conventional way can easily be thrown off by an unfamiliar object or a slight variation in lighting.OpenAI confirmed that it is working with the robots from Fetch, but it declined to comment further. Melonee Wise, the company’s founder, couldn’t be reached for comment.OpenAI was created by Musk and a handful of well-known (and well-heeled) Silicon Valley entrepreneurs, including investor Peter Thiel, Y Combinator president Sam Altman, and the incubator’s cofounder Jessica Livingston. The nonprofit’s backers have committed $1 billion in funding to the project, and it is being led by Ilya Sutskever, a prominent AI researcher who left Google to join the project, and Greg Brockman, an early employee at the high-profile digital payment company Stripe.While OpenAI has committed to making the technology it develops publicly available, it could certainly benefit companies backed by Musk and Thiel, as well as those emerging from Y Combinator.Produced by Will Knight 

Engineers at the University of California San Diego have developed a flexible wearable sensor that can accurately measure a person's blood alcohol level from sweat and transmit the data wirelessly to a laptop, smartphone or other mobile device. The device can be worn on the skin and could be used by doctors and police officers for continuous, non-invasive and real-time monitoring of blood alcohol content.The device consists of a temporary tattoo—which sticks to the skin, induces sweat and electrochemically detects the alcohol level—and a portable flexible electronic circuit board, which is connected to the tattoo by a magnet and can communicate the information to a mobile device via Bluetooth. The work, led by nanoengineering professor Joseph Wang and electrical engineering professor Patrick Mercier, both at UC San Diego, was published recently in the journal ACS Sensors."Lots of accidents on the road are caused by drunk driving. This technology provides an accurate, convenient and quick way to monitor alcohol consumption to help prevent people from driving while intoxicated," Wang said. The device could be integrated with a car's alcohol ignition interlocks, or friends could use it to check up on each other before handing over the car keys, he added."When you're out at a party or at a bar, this sensor could send alerts to your phone to let you know how much you've been drinking," said Jayoung Kim, a materials science and engineering PhD student in Wang's group and one of the paper's co-first authors.Blood alcohol concentration is the most accurate indicator of a person's alcohol level, but measuring it requires pricking a finger. Breathalyzers, which are the most commonly used devices to indirectly estimate blood alcohol concentration, are non-invasive, but they can give false readouts. For example, the alcohol level detected in a person's breath right after taking a drink would typically appear higher than that person's actual blood alcohol concentration. A person could also fool a breathalyzer into detecting a lower alcohol level by using mouthwash.Recent research has shown that blood alcohol concentration can also be estimated by measuring alcohol levels in what's called insensible sweat—perspiration that happens before it's perceived as moisture on the skin. But this measurement can be up to two hours behind the actual blood alcohol reading. On the other hand, the alcohol level in sensible sweat—the sweat that's typically seen—is a better real-time indicator of the blood alcohol concentration, but so far the systems that can measure this are neither portable nor fit for wearing on the body.Now, UC San Diego researchers have developed an alcohol sensor that's wearable, portable and could accurately monitor alcohol level in sweat within 15 minutes."What's also innovative about this technology is that the wearer doesn't need to be exercising or sweating already. The user can put on the patch and within a few minutes get a reading that's well correlated to his or her blood alcohol concentration. Such a device hasn't been available until now," Mercier said.How it worksWang and Mercier, the director and co-director, respectively, of the UC San Diego Center for Wearable Sensors, collaborated to develop the device. Wang's group fabricated the tattoo, equipped with screen-printed electrodes and a small hydrogel patch containing pilocarpine, a drug that passes through the skin and induces sweat.Mercier's group developed the printed flexible electronic circuit board that powers the tattoo and can communicate wirelessly with a mobile device. His team also developed the magnetic connector that attaches the electronic circuit board to the tattoo, as well as the device's phone app."This device can use a Bluetooth connection, which is something a breathalyzer can't do. We've found a way to make the electronics portable and wireless, which are important for practical, real-life use," said Somayeh Imani, an electrical engineering PhD student in Mercier's lab and a co-first author on the paper.The tattoo works first by releasing pilocarpine to induce sweat. Then, the sweat comes into contact with an electrode coated with alcohol oxidase, an enzyme that selectively reacts with alcohol to generate hydrogen peroxide, which is electrochemically detected. That information is sent to the electronic circuit board as electrical signals. The data are communicated wirelessly to a mobile device.Putting the tattoo to the testResearchers tested the alcohol sensor on 9 healthy volunteers who wore the tattoo on their arms before and after consuming an alcoholic beverage (either a bottle of beer or glass of red wine). The readouts accurately reflected the wearers' blood alcohol concentrations.The device also gave accurate readouts even after repeated bending and shaking. This shows that the sensor won't be affected by the wearer's movements, researchers said.As a next step, the team is developing a device that could continuously monitor alcohol levels for 24 hours. Provided by: University of California - San Diego