Robots that can hold simple conversations and learn people's interests are to be deployed in some UK care homes after an international trial found they boosted mental health and reduced loneliness. The wheeled robots, called "Pepper", move independently and gesture with robotic arms and hands and are designed to be "culturally competent", which means that after some initial programming they learn about the interests and backgrounds of care home residents. This allows them to initiate rudimentary conversations, play residents' favourite music, teach them languages, and offer practical help including medicine reminders. The researchers, led by Dr Chris Papadopoulos at the University of Bedfordshire, said the trial was not intended to explore the replacement of human carers with robots, but to help fill lonely periods when, because of a stretched social care system, staff do not have time to keep residents company. The trial, in the UK and Japan, found that older adults in care homes who interacted with the robots for up to 18 hours across two weeks had a significant improvement in their mental health.

HSBC is replacing more manual processes, with artificial intelligence (AI) being used to automate when ATMs need to be refilled. The technology, developed by HSBC's operations and technology teams, has been trialled in Hong Kong, where the bank has 1,200 ATMs. The iCash AI technology has reduced ATM refills, which are done by third parties, by 15% – saving $1m. To calculate how much money is needed and where, iCash uses live ATM data and predictive machine learning algorithms that factor in seasonality, holidays, public events, location and recent withdrawal trends. The bank said it was a challenge to predict how much cash each ATM might need.

In mid-July, a UPS subsidiary called Flight Forward and the drone company Matternet started a project with the Wake Forest Baptist Health system in North Carolina. The companies' aims are decidedly futuristic: to ferry specialty medicines and protective equipment between two of the system's facilities, less than a half-mile apart. Think of it: little flying machines, zipping about at speeds up to 43 mph, bearing the goods to heal. At this point, though, the drone operations are a little, well, human. The quadcopters must be operated by specialized drone pilots, who must pass a challenging aeronautical knowledge test to get their licenses.

Yandex is basically the Google of Russia. Russian technology company Yandex has been working on self-driving vehicles since 2017. Similarly, it partnered with American firm Uber to form a ridesharing and food-delivery joint-venture. On Friday, the two companies announced they're spinning the autonomous-vehicle portion of the business off as a separate entity. Once the financial dust settles, the unimaginatively named Yandex Self Driving Group, or SDG, will be directly owned by both businesses, with Yandex holding about 73% of SDG and Uber around 19%.

With recent advances, the tech industry is leaving the confines of narrow artificial intelligence (AI) and entering a twilight zone, an ill-defined area between narrow and general AI. To date, all the capabilities attributed to machine learning and AI have been in the category of narrow AI. No matter how sophisticated – from insurance rating to fraud detection to manufacturing quality control and aerial dogfights or even aiding with nuclear fission research – each algorithm has only been able to meet a single purpose. This means a couple of things: 1) an algorithm designed to do one thing (say, identify objects) cannot be used for anything else (play a video game, for example), and 2) anything one algorithm "learns" cannot be effectively transferred to another algorithm designed to fulfill a different specific purpose. For example, AlphaGO, the algorithm that outperformed the human world champion at the game of Go, cannot play other games, despite those games being much simpler.

Rapid prototyping, the fast and cost-effective production of prototypes or models -- better known as 3D printing -- has long since established itself as an important tool for industry. "If this concept could be transferred to the nanoscale to allow individual molecules to be specifically put together or separated again just like LEGO bricks, the possibilities would be almost endless, given that there are around 1060 conceivable types of molecule," explains Dr. Christian Wagner, head of the ERC working group on molecular manipulation at Forschungszentrum Jülich. There is one problem, however. Although the scanning tunnelling microscope is a useful tool for shifting individual molecules back and forth, a special custom "recipe" is always required in order to guide the tip of the microscope to arrange molecules spatially in a targeted manner. This recipe can neither be calculated, nor deduced by intuition -- the mechanics on the nanoscale are simply too variable and complex.

The design of real-world automated control systems that do everything from regulating the temperature of skyscrapers to running the widget-making machine in the widget factory down the street requires expertise in sophisticated physics-based modeling. The need for this modeling expertise increases operational costs and restricts the applicability of automated control to systems in which marginal operational performance improvements lead to huge economic benefits, according to data scientists. With unlimited access to supercomputers and mountains of data, engineers can train artificial intelligence systems such as deep neural networks, a type of machine learning model, to perform automated control. But many people lack access to the necessary computational power to do so, or the ability to generate the amount of data needed to train a controller that has a deep neural network. What's more, these types of deep neural networks are so-called black-box models, which means that the factors they use to make decisions are hidden from the end user.

The next layer embodies the semi autonomous weapons which despite lifting some of the risk off the human operator, they can't assume full responsibility since the human operator still has the final say. Examples of those kind of weapons are the fire-and-forget air-to-air missiles or aircraft which automatically lock onto their targets. Pressing the button to launch such missiles is still the pilot's responsibility... Clearly the danger lurks in the third level, that of the fully autonomous robots where no human intervention is required. These robots are able to act on their own and accomplish their mission without the emotional burden or ethical reservations there to stop them.

Bezos said in the 2013 interview that it would take four or five years to have those drone deliveries. It turns out that using remote-controlled aerial gizmos to drop stuff at our homes is incredibly difficult, prone to risk and potentially more trouble than it's worth. Like driverless cars, drone technology in populated areas is more complicated than most people expected, and it has been -- mostly for good reason -- tightly controlled in the United States by government agencies worried about drones straying into the path of airplanes, dropping out of the sky onto our heads or unwittingly spying through people's windows. It wasn't until this week that the F.A.A. gave Amazon permission to do drone deliveries. And drones might never be practical for deliveries when someone in a vehicle could do the same thing in a fraction of the time and cost. Drones are a great public relations jolt for Amazon, but let's not put too much stock in them for awhile -- maybe ever.

Recently, researchers from DeepMind and McGill University proposed new approaches to speed up the solution of complex reinforcement learning problems. They mainly introduced a divide and conquer approach to reinforcement learning (RL), which is combined with deep learning to scale up the potentials of the agents. For a few years now, reinforcement learning has been providing a conceptual framework in order to address several fundamental problems. This algorithm has been utilised in several applications, such as to model robots, simulate artificial limbs, developing self-driving cars, play games like poker, Go, and more. Also, the recent combination of reinforcement learning with deep learning added several impressive achievements and is found to be a promising approach to tackle important sequential decision-making problems that are currently intractable.