The DURUS robot can walk more than a mile in man's shoes. A pair of size 13, Adidas sneakers, to be specific. Engineers at the Georgia Institute of Technology have tackled what they describe as a deceptively difficult challenge: develop a battery-powered robot that mimics the subtle complexity of the human footstep. Aaron Ames, an associate professor of automation and mechatronics, said their feat represents a stride in robotic efficiency and mobility and could allow for robots to function more seamlessly in environments meant for humans. "What drives me a lot is the cool factor, to be honest, but that's my professor hat," Ames said.

Many problems in science and engineering can be formulated as a sequential decision-making problem under uncertainty. A common scenario in such problems that occurs in many different fields, such as online marketing, inventory control, health informatics, and computational finance, is to find a good or an optimal strategy/policy, given a batch of data generated by the current strategy of the company (hospital, investor). Although there are many techniques to find a good policy given a batch of data, only a few of them guarantee that the obtained policy will perform well, when it is deployed. Since deploying an untested policy can be risky for the business, the product (hospital, investment) manager does not usually allow it to happen, unless we provide her/him with some performance guarantees of the obtained strategy, in comparison to the baseline policy (e.g., the policy that is currently in use). In this paper, we focus on the model-based approach to this fundamental problem in the context of infinite-horizon discounted Markov decision processes (MDPs). In this approach, we use the batch of data and build a model or a simulator that approximates the true behavior of the dynamical system, together with an error function that captures the accuracy of the model at each state of the system. Our goal is to compute a safe policy, i.e., a policy that is guaranteed to perform at least as well as the baseline strategy, using the simulator and error function. Most of the work on this topic has been in the model-free setting, where safe policies are computed directly from the batch of data, without building an explicit model of the system [12, 13]. Another class of model-free algorithms are those that use a batch of data generated by the current policy and return a policy that is guaranteed to perform better.

With the JIMU robot, UBTECH Robotics, a Schenzen-based company that has been around for eight years, is stepping into a space that's quickly getting cluttered with motorized toys. The company already has a lineup that includes industrial bots in China and commercial humanoids like Alpha 2, but now they're using their existing infrastructure to build affordable, programmable robots for kids, eight and older, with DIY inclinations. They might also find a home in schools that are looking to adopt coding in their curriculum. The MeeBot kit, which is available exclusively at Apple stores for 130 starting today, comes with interlocking parts that include colorful blocks, connectors, motors and a rechargeable lithium-ion battery. A platinum grey control box with a U-shaped line across the front doubles as a smiling face.

Their squishy robotic hands can gather coral samples more delicately than robots, and in places humans can't reach. Developed with support from a National Geographic Innovation Challenge Grant, the hands were first tested in tanks in March 2015 and then taken to the Red Sea in May. After a successful expedition, Wood and Gruber hope the technology may have even broader applications.

For aerial robots, maintaining a high vantage point for an extended time is crucial in many applications. However, available on-board power and mechanical fatigue constrain their flight time, especially for smaller, battery-powered aircraft. Perching on elevated structures is a biologically inspired approach to overcome these limitations. Previous perching robots have required specific material properties for the landing sites, such as surface asperities for spines, or ferromagnetism. We describe a switchable electroadhesive that enables controlled perching and detachment on nearly any material while requiring approximately three orders of magnitude less power than required to sustain flight. These electroadhesives are designed, characterized, and used to demonstrate a flying robotic insect able to robustly perch on a wide range of materials, including glass, wood, and a natural leaf.

Flapping two tiny wings, the small, thin robot wobbles its way towards the underside of a leaf, bumps into the surface and latches on, perching motionless above the ground. Moments later, its wings begin to flap once more and it jiggles off on its way. The little flying machine, dubbed a "RoboBee", has been designed to perch on a host of different surfaces, opening up new possibilities for the use of drones in providing a bird's-eye view of the world, scientists say. Know as micro aerial vehicles, such robots could be invaluable in reconnaissance of disaster zones or to form impromptu communication networks. But there is a hitch: flying takes energy, so the time these robots can spend in the air is limited by the size of the battery pack they can carry.

The promise of virtual reality is that it can transport you to places you'd prefer not to go: the tops of the highest mountains, for instance, or the mosh pit of Norwegian party metal concert. Then there are the impossible places, like the roiling vents at the bottom of the deepest oceans, where crushing pressures and searing heat make an environment fit only for robots. In March, one of those robots, the straightforwardly-named Remotely Operated Vehicle for Ocean Sciences, spent a staggering 150 hours exploring an undersea volcano near Samoa. Not only were researchers from the Schmidt Ocean Institute able to capture VR video and upload it to YouTube so regular folk can explore the action themselves (check it out below), but they 3-D mapped a so-called black smoker vent so scientists around the world can study the phenomenon independently. But this wasn't all an exercise in delayed gratification.

The expedition has only just begun and already it is producing some stunning footage of the denizens of Earth's deepest abyss. NOAA's Okeanos research ship will spend the next two months cruising the seas above the Mariana trench, using sonar systems and a remotely operated vehicle (ROV) to explore the little known parts of the trench and its surroundings. Footage from the daily dives is being live-streamed, with highlights posted on the expedition site. On 24 April the ROV captured this footage of a beautiful jellyfish (see video above), belonging to the Crossota genus, at a depth of nearly 4 kilometres. On 25 April it filmed a striking, violet sea cucumber swimming above the sea floor (see video below).

A recent article titled "Why is AI Female?" made the connection that gendered labor, in service professions in particular, is fueling our expectations for gendered AI assistants and service robots. Furthermore, the author argues, this "feminizing -- and sexualizing -- of machines" signals a future with a disproportionate use of feminized VR and robots for a male-dominated sex industry. "Sex with robots is a big leap from asking Siri to set an alarm, but the fact that we've largely equated artificial intelligence with female personalities is worth examining. There are, after all, few sexualized male robots or avatars." Herbert Televox and Mr. Telelux, the early 20th century robots made by Westinghouse, were both male.

A recent article titled "Why is AI Female?" made the connection that gendered labor, in service professions in particular, is fueling our expectations for gendered AI assistants and service robots. Furthermore, the author argues, this "feminizing -- and sexualizing -- of machines" signals a future with a disproportionate use of feminized VR and robots for a male-dominated sex industry. "Sex with robots is a big leap from asking Siri to set an alarm, but the fact that we've largely equated artificial intelligence with female personalities is worth examining. There are, after all, few sexualized male robots or avatars." Herbert Televox and Mr. Telelux, the early 20th century robots made by Westinghouse, were both male.