Will YouTube revolutionize research?

In September 2011 at the University of California, Berkeley, an experiment was conducted to scan the brain of a person watching a movie. The work was conducted by prof. Jack Gallant. By brain scan, it was possible to accurately visualize exactly small parts of the brain along with its activity. The exact definition of what part of the brain is active allows you to determine what shape it may actually correspond to. These data were transferred to a computer that compiled them into a database of 18 million 1-second videos of specific YouTube objects.

The most accurate "hits" was selected and incorporated into a larger film projection so that it allowed it to recreate what the brain sees (and in principle thinks). Generated images are unfortunately very blurred. This is because the computer imposes a second movie at a very fast pace, but still much slower than the brain works. Depending on the specific shape, some of the generated videos may surprise you with similarity. It's hard to say here about the milestone.

This, though, a fairly accurate experiment, however, offers tremendous opportunities and shows that theoretical assumptions work in practice. One can hope that humanity will soon be able to portray a dream film, brain vision or even his thoughts. The only "need" for a computer that would think as fast as a brain would process data in relation to the human brain at a 1: 1 scale.

.

Currently the fastest computer on earth, so called. Fujitsu K Computer, has a computing power of 10.5 PFLOPS at 1 petabyte RAM. Such performance enabled the creation of a 1 second simulation of the activity of 1.74 billion nerve cells connected by 10.4 trillion synapses. The time to produce this one-second simulation is ... 40 minutes. The number of nerve cells whose work is stimulated is only 2% of the whole neural network of the human brain! So to computer reconstruction of the work of the human brain and to read his thoughts still a long way. Hope may be quantum computers.

Breakthrough in 3 years?

Recently in the prestigious magazine Nature, Chinese scientists from the Chinese University of Science and Technology have announced the breakthrough in the construction of this device. Their prototype is several dozen times faster than the first electronic computer constructed in the 1940s. The researchers took up the so-called cubits, the basic units of information. In a classic computer the information is written in the form of strings of zeros and ones. Meanwhile, kubit can be zero and one at the same time, which should give scientists the ability to write different information in the same memory cells at the same time. However, this requires the development of a special, quantum control program.

Zhu Xiaobo, a member of the band, reported that his team broke the record. "In order for quantum computers to work well, we need to pair all the cubits with us. This is called confusion. And we have a record number of superconducting cubits, that is ten, "he explained in a publication.

Scientists, not just those of you, argue that while work on quantum computers is slower than expected, around 2020 we can expect these devices to enter IT systems.