innovator

Add idea


Calendar

«    July 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
31
 

 

Advert

 

Payment

 

Advert

 

Authorization

Стартап

DateDate: Today, 06:25

The technology of classical convolutional neural networks - or as they are also called cCNN - is the main base of algorithms and scripts, thanks to which various neural networks can analyze visual information, correctly process it and fill in the missing fragments. Being a technology that is fairly common at all levels, it nevertheless has a number of significant limitations that do not allow it to be called truly effective and useful in applying it in really complex tasks. For example, in the processing of visual information for artificial intelligence of autonomous cars or forensic neural networks.
It is for this reason that Till Hartman, a talented neuroscience researcher who is a professor of neuroscience at Harvard University, decided to somewhat improve this neural network technology, making it more human-like. In the sense to teach it to think and visualize the space in the way that the human brain does - after all, filling in the missing visual information often takes place under imperfect and often problematic conditions.
He drew attention to how the visual cortex functions, noting its reverse-direct type of visualization and processing of visual information. Joining with the software engineering team, he managed to present so-called recurrent convolutional neural networks that almost completely repeat the process of processing and perceiving the visual information of the human brain, thus filling the empty logical space of the existing display much more efficiently.
Moreover, these CNN reflexes have a truly ambitious level of work, even under the most disturbing circumstances and conditions - for example, the new system of neural networks was tested to determine how well it can complement the blurred image, which it coped well after a few seconds of analysis.

DateDate: Yesterday, 07:21

Question about Apple's own car is still open.
Apple developed its own suspension patent for the first time in the United States Patent and Trademark Office two years ago, but the USPTO just published the documents and made them public on Tuesday.
The language of the patent is full of engineering slang, but the basic idea is in this configuration of the suspension, which will receive a special component connecting the frame and the steering wheel.
Such a suspension can be used to provide the most comfortable steering, including in an autonomous car. In addition, such a system, according to Apple, will reduce tire wear and increase vehicle stability.

DateDate: 18-07-2019, 05:54

Ilon Mask presented neural interfaces for connecting the human brain to computers, reports the New York Post.
The device is a "thread" that can be embedded in the brain to transfer large amounts of data to a computer.
According to representatives of the company Ilon Mask Neuralink, about 20 rodent operations have already been performed, 87% of which were successful.
The main purpose of the device is to implant it to paralyzed people so that they can remotely control computers and telephones.
Earlier, the "Journalist" reported that Ilon Musk showed the first photo of the new Starship spacecraft.

DateDate: 17-07-2019, 05:57

The Hong Kong company introduced a rescue buoy with a radio-controlled motor.
A rescue buoy on the radio has been developed.
The Hong Kong company OceanAlpha has developed Dolphin 1, a rescue radio-controlled buoy. It can move through the water at a speed of up to 15 kilometers per hour up to a distance of half a kilometer from the operator. This is stated on their official website.
If Dolphin rolls on the waves, the manager can flip it back without getting it out of the water. The engines are insulated with a plastic cap, which prevents algae and debris from entering the propeller, and also ensures the safety of people during transportation.
Dolphin 1 is on sale, one device will cost 5.5 thousand dollars.

DateDate: 16-07-2019, 05:57

A special thermal battery will help solar stations generate energy at night. In addition, it will provide an opportunity to store electricity.
Thermal Battery, United Sun Systems, ITP Thermal, Solar Energy
It is reported Sotsportal with reference to Hightech.
- The thermal battery was developed by specialists from Curtin University and international companies United Sun Systems and ITP Thermal as part of a project to create a system of concentrated solar energy, which requires a battery. It provides an opportunity to store energy and give it away at any time of the day, seven days a week, writes Phys.org. The ultimate goal of the project is the development of a new technology that integrates the thermochemical energy storage into the parabolic Stirling concentrator. This technology produces up to 46 kW of energy and is ideal for remotely located industrial enterprises, such as mines, because it provides them with the right amount of electricity on demand. Depending on your needs, you can install several parabolic plates.
A thermal battery developed by a team of Australian scientists will become part of this power grid. It accumulates heat in the metal, and gaseous hydrogen or carbon dioxide is used as the coolant. At night, or when it is overcast, hydrogen or carbon dioxide from the tank is absorbed by the warmer metal, forming a metal hydride or carbonate that releases heat. Then heat is converted into electricity using a thermogenerator.
“Energy storage has always been a stumbling block for renewable energy, but our prototype thermal battery can store and, if necessary, give off solar energy regardless of sunlight,” explained Professor Craig Buckley, project manager.