Add idea


«    March 2019    »










DateDate: 15-02-2019, 06:18
A train longer than 30 meters can accommodate about 300 passengers.
The train on rubber tires can move on any road. With the help of built-in sensors automatically plot the route taking into account the qualities of the road.
The maximum train speed is 70 km / h, after a 10-minute charge it can travel about 25 kilometers.
Also, earlier it was reported that the Chinese company Star UAV System has started mass production of the AT200 transport unmanned aerial vehicle.
Tests of an unmanned vehicle were carried out in 2017-2018 and were recognized as successful, after which the developers started production of the AT200. Interest in the new apparatus has already shown the command of the People’s Liberation Army of China.

DateDate: 14-02-2019, 05:57

The Beijing Architecture School at Tsinghua University printed a 26.3-meter-long bridge from a concrete printer on a 3D printer, reports with reference to Dezeen.
It is alleged that a pedestrian bridge with a width of 3.6 meters, passing through a canal in the Baoshan area, has become the longest of its kind in the world. The authors of the project repeated the arched form of the oldest, 1400-year-old Chinese Anji bridge.
According to the designers, printing a bridge turned out to be a third cheaper compared to traditional construction. It was possible to save due to the lack of fittings and crates.
The single-arch bridge is constructed of polyethylene fiber-reinforced concrete with additives additives. It consists of 176 concrete blocks that were printed using two 3D printing systems with a robotic arm. It took 450 hours to print all the elements.
A monitoring system is built into the bridge that will record the deformation of the concrete over time.

DateDate: 13-02-2019, 07:09

At the World Economic Forum in Davos, Skaterbot was presented - a four-legged robot developed at the Swiss Higher Technical School (Zurich). Instead of wheels on which already known walkers move, Skaterbot has attached skates, writes National Geographic.
When the robot made the first movements on the ice, one of the spectators exclaimed: “I started the same way!”
Skaterbot is able to adapt to the surface and, thus, learn to move on it independently.
The video shows that the legs of the development move alternately to the right and left, which allows it to continue moving along the slippery surface. In just a few hours of training, Skaterbot developed a gait in which his legs constantly turned left and right, creating undulating movements.
In addition to skates, Skaterbot can adapt to different formats of the structure - the engineers have already taught him to move on a walking chassis, as well as on tracks.
Recall that the Swiss Federal Institute of Technology introduced a new version of the four-legged robots ANYmal - the device was taught to move faster than ever, and not let yourself be knocked down.

DateDate: 12-02-2019, 06:24

Artificial intelligence can play chess, drive cars and look for new medicines. This type of AI is called limited or weak. These are just advanced algorithms that perform specific tasks. Their abilities and scope of application is constantly growing, but their danger to humans is limited by the loss of work, reports HighTech.
The new generation of AI will certainly raise the stakes, scientists say. The combination of computing power of machines and human intelligence will allow a strong AI to learn, solve complex tasks and improve themselves. And to engage in tasks for which no one prepared them. So AI can turn into an artificial super-intelligence (ICI). According to some experts, its appearance is possible in the period from 2029 to the end of the century, according to Science Alert.
Whenever this happens, there is no doubt that the ISI will change humanity. Perhaps he will solve the world's problems of lack of resources or climate change.
But the lack of proper control will lead to disastrous results. And it's not about killer robots.
IISI is unlikely to be hostile to people. But the car can come to the conclusion that the most effective strategy to solve the problem that people have assigned to it is to get rid of some people.
Here are the possible scenarios. ISI, which develops a cure for the virus, decides to kill all its carriers. The military drone comes to the conclusion that the only way to ensure the security of the allied forces is to destroy the entire settlement along with the civilians. A neural network, designed to protect humanity from climate catastrophe, erases from the face of the Earth all the technologies that pollute it.
Now, such layouts seem fantastic, but even on a more modest scale, they can cause a lot of harm. To avoid such scenarios, scientists propose to introduce three methods of control.

DateDate: 11-02-2019, 07:01

Police departments in various cities in the United States and Great Britain secretly signed contracts with PredPol, the developer of an ambiguous and untested offense prediction program.
This fact was noticed by Motherboard journalists, who analyzed documents received on the basis of archival requests, in particular, contracts, instructions and program presentations, as well as emails discussing the terms of the contract between representatives of PredPol and heads of police departments in at least 15 cities and municipalities of the United States and Great Britain.
PredPol claims that it uses an algorithm to calculate the probability of crime in an urban area measuring 150 × 150 meters.
Predictions are compiled on the basis of statistics on the crimes committed in the same place and at the same time. The forecast can be obtained for the next 3, 7, 14 or 28 days. Based on this information, a police patrol can be sent there in the hope that the system was not mistaken and that it would be necessary to prevent the offense.
“We use data on crimes for 3 - 10 years and run them through our algorithm,” explains the representatives of ProdPol. “Short-term and long-term trends, relapses and environmental factors are all taken into account.”
Prevention also helps potential offenders. “They have another opportunity to avoid conflict with the law enforcement system,” ProdPol says.
Criminologists, however, believe that this approach does not guarantee more effective patrolling of areas. Shahid Buttar, director of the human rights organization Electronic Frontier Foundation (EFF), believes that the predictive technologies are inevitably biased, because such are the data with which they work.
If you catch more police into one area, more crimes will be found there. Then the AI predicts that this is a very criminogenic place.
Andrew Ferguson, a law professor at the University of the District of Columbia, stresses: we don’t know if the algorithm helps reduce crime. Objective scientific data are too few, there are no studies on this topic. “We are talking about the police, and people in the USA die at the hands of the police every day,” he said.
In the autumn, it became known that Kent County police terminated the contract with PredPol, which cost taxpayers $ 128,000 a year. They decided that the algorithm, although predicting crimes, did not help to prevent them very effectively. However, the idea of the police like, not like only the price. So now they have undertaken to independently develop similar software.