innovator

Add idea


Calendar

«    April 2019    »
MonTueWedThuFriSatSun
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
 

 

Advert

 

Payment

 

Advert

 

Authorization

Стартап

DateDate: 14-04-2019, 07:28

Today, getting from point A to point B is not so difficult. Even if you do not know the city or area - one of the many navigation systems or the same maps from Google will tell you the way. Now imagine a situation in which, for sure, many turned out to be: you got to the destination, enter the building and ... get lost. Where to go? Where is my office / store / department? It is quite possible that such situations will soon be avoided, because the US company is developing a technology for mapping and indoor navigation based on augmented reality (AR).
It is worth noting that Apple already promised something similar to us. Only here the Cupertini built their system on iBeacon devices that need to be installed. Immediately, no "beacons" are needed. Only a smartphone is needed.
Everything works quite simply: the technology is integrated into the internal maps of the premises. Thanks to this, she can find out where the user is at the moment, and then show the right directions on the smartphone screen. To do this, use the camera of the device, which "removes" the room, and the path is projected on the screen in real time. In fact, this is very similar to the AR mode in Google Maps, except that it is intended for indoor use and not for street navigation.
The development is called Dent Reality and the software part relies on both GPS data and special algorithms for determining the position of the user. Algorithms include methods for measuring distance, recognizing objects inside a building, and options for displaying a virtual path with incredible accuracy.
Moreover, the program is based on an open SDK, so third-party developers can make their own improvements in it, as well as use the software part to develop their products of a similar theme. In addition, the developers have opened access to the materials on the basis of which the algorithms were built.
At the moment, the technology looks very promising, but it is at a very early stage and still far from release. However, if the developers and their partners succeed, we will get a very convenient tool for orientation in unfamiliar rooms.
Source: androidinsider.ru

DateDate: 13-04-2019, 06:39

The device is based on the principle of absorption and heat during deformation of the metal.
A group of German researchers at Saarland University (Saarland University) has created an unusual heating-cooling system, which, when used in refrigeration, can at least be three times more efficient than modern similar devices, not to mention its other advantages, reports Chronicle.info with referring to Esoreiter.
If we speak in a somewhat simplified language, then the essence of this refrigeration generator lies in a rotating cylinder, on which the nitinol wire is wound, which stretches in one direction of rotation, and in another is compressed. Nitinol (a unique alloy) is capable of absorbing heat when it is expanded, and when it returns to its original position, it can give off. Experiments have shown that the temperature difference in this case can reach, even at the minimum value, 20 degrees.
This is what one of the co-authors of this invention, Professor Stefan Seeleke, says about this:
We were able to calculate the diameter and size of the nitinol wire, the speed of the drum, the parameters of the air flow from the fan and so on, as a result of which we achieved not only an impressive temperature difference of 20 degrees Celsius, but also the system power 30 times greater than the effort required by the operation of this installation.
It is three times more efficient than a modern refrigerator. But in principle, in the future when creating industrial units of this type, the indicators can be improved several times. Moreover, this refrigerator does not consume toxic substances (environmentally friendly), does not require hermetic pipes, heat exchangers and other intermediate links.
See also: The man told what he saw during clinical death
There is only one very important question: if the nitinol wire will constantly expand and shrink again, how much is enough? But it turns out that this alloy differs from many others in its fantastic fatigue resistance, which makes nitinol, for example, used in arterial stents, which allows such an implant to shrink, curl, stretch and so on along with the blood vessel for an infinitely long time. In addition, this metal is relatively cheap, which, together with countless expansion-compression cycles, makes it optimal for the cooling system in question.

DateDate: 12-04-2019, 05:56

In Switzerland, completed the first house, which was built by robots, designed computers, and parts made 3D-printers. So far, however, with the participation of man. In Switzerland, they put into operation a house that was built by Digital Manufacturing and Living robots - this is how its name (Digital Fabrication and Living) stands for. DFAB HOUSE is a joint demonstration of the digital production of the Swiss National Center for Competence in Research (NCCR) in the NEST (Nest) building. As part of a full-featured construction project, researchers from Zurich (ETH Zurich), Switzerland, have gathered together with industry experts and planners to explore and test how digital production can change our design and construction methods.
The “house on the roof” space is planned to be used not only as housing, but also as a testing ground, on which construction and technological innovations will be tested in real conditions. Within this project, for the first time, six new digital construction processes were transferred from research to architectural applications:
In situ Fabricator, the universal autonomous building robot;
Mesh Mold, a non-formwork robotic process for reinforced concrete structures;
Smart Dynamic Casting, an automated concrete molding process;
Smart Slab, built-in ceiling tiles made with 3D-printed formwork;
Spatial assembly of wooden structures;
Robotic construction of wood.
Combining these new processes in a single constructed facility allows you to rethink the entire planning and construction process and use the advantages inherent in the digital chain of design, planning and manufacturing: design flexibility, saving material, saving time and cost and improving quality. It is already clear that the architectural potential of digital technology is enormous, but is almost never used on construction sites. Experimental projects like DFAB should speed up the transition from theory to practice, says ETH Zurich professor Matthias Kohler. The house turned out to be futuristic: the blinds are raised on command and the water in the kettle starts to boil; It has a multi-stage security and lighting control system.
The work of the smart home is provided by digitalSTROM equipment. The “smart” technologies of this house work not only at the household level, they also help to control energy consumption. Solar cells on the roof provide energy - on average, one and a half times more than is necessary to maintain the house, and the “intelligent” control system controls its consumption and smoothes the load peaks. That is, it is an "active home". Heat from wastewater is not wasted, but transferred further through heat exchangers installed in shower trays. Unused hot water is returned through the pipes back to the boiler, which allows not only to save energy and water, but also to prevent the growth of bacteria in the pipes.

DateDate: 11-04-2019, 05:56

The academic publishing house Springer Nature presented the first research book created using machine learning. Reported by The Verge.
The book "Lithium-Ion Batteries: A Machine-Generated Summary of Current Research" contains a summary of peer-reviewed articles published on this topic, including citations, hyperlinks and automatically generated links to the content. The tutorial is available for free download.
In his opening remarks, Henning Schoenenberger of Springer Nature said that such books are capable of starting "a new era in scientific publications, automating routine work."
Schoenberger notes that in the last three years alone more than 53,000 scientific papers on lithium-ion batteries have been published. This is a huge problem for scientists who are trying to keep abreast of developments. But using AI to automatically scan and summarize these results, scientists can save time and continue important research.
"This method allows readers to speed up the process of mastering the literature in this field of research. At the same time, if necessary, readers can always identify the source and go to it for further study of the subject," said Schönenberger.
1.png (83 KB)
Although the recent boom in machine learning has greatly improved the ability of computers to generate the written word, the performance of these bots is still very limited. They cannot struggle with the long-term consistency and structure that human writers create, and therefore such efforts as fiction or poetry generated by AI are more likely to be formatted.
What an AI can do is compose formula texts when loading a library. For example, in journalism, machine learning is used by organizations such as The Associated Press to create reports of football matches, earthquakes, and financial news. These are topics where creativity is an obstacle.
According to associate professor at the Institute of Human Computer Interaction Jeff Bigham, the book was not the most impressive feat of AI. “It’s enough just to take a high-quality introductory text, briefly summarize the main conclusions and make the material look connected. In fact, the very nature of the extract means that it will be consistent in parts if the input texts are coherent. would call it valuable, "commented Bigham.
Earlier, InternetUA has already reported on AI journalism: the world's largest media uses robot authors to create articles. Moreover, AI journalists are already writing fake news - computers that have already been used to create short news reports from press releases can be taught to read and write long false texts.

DateDate: 10-04-2019, 05:55

Scientists have conducted experiments to test how much artificial intelligence is smarter than man.
As an experimental, the researchers took the DeepMind neural network and decided to offer it to independently examine the school curriculum in mathematics, and also to perform tasks from the tenth grade level. Reports the edition about it "Today".
As a result, DeepMind failed the school test with disgrace, gaining 14 points out of 40. Most British students receive marks much higher.
Artificial intelligence was unable to adequately evaluate test assignments. If an ordinary schoolchild can easily recognize a huge number of numbers, letters and formulas, then this is given to a neural network with difficulty. Most of the time, DeepMind spent on character recognition, as well as translating information into a readable form for its own analysis spent a lot of energy.
Meizu got into a big scandal: what happened
Meizu got into a big scandal: what happened
technology
But still, artificial intelligence is able to solve a complex mathematical problem. For this, it must be presented in a special form so that the neural network does not spend its resources on recognition. In the meantime, the "visual" information is processed by neural networks with difficulty, which promises such distressing results.
Earlier it was reported that artificial intelligence has learned to predict death to the nearest second.
Thus, the case histories of half a million Britons aged 40 to 69 years were chosen as the material for the study.
Artificial Intelligence
Researchers from the University of Nottingham began testing a self-learning algorithm that they had independently developed. This algorithm is able to assess the risks of premature death of a person suffering from chronic diseases. And also knows how to name almost the exact date of departure from a person’s life.
Half a million Britons aged 40 to 69 years of age were chosen for the study.
After comparing the data, it turned out that the neural network predicted human deaths more accurately than the system used by humans.