Simulations of the brain

Computer technology is more developed in the direction of role models of biological processes, at first — thinking, and this immediately leads to a mind-blowing theories until the artificial mind.

"I'm smart enough … And the average person … such weak … kopotlivy … "- these words are pronounced in Asimov's robot rebellious. There is a picture of the imperfection of the human brain in relation to the computer. The situation is, in fact, the reverse — it's computers at the moment are trying to build in imitation of the brain, and the proof of this — messages of different research groups made in recent weeks.

1014 — the so-called report on the progress of the project COMPASS at IBM. 10 in the fourteenth degree — is the advance estimate the number of bonds between the cells of a human brain, and specifically that the order has been made in recent computer simulations of the project. "Neurons" and "synapses" (intercellular connections of neurons) — the dictionary definition of the physiology of the brain — firmly established themselves in the field of computer engineering.

Simulations of the brain

Architecture conventional computers we made even in the middle of the last century and named "Von Neumann" in honor of the mathematician John von Neumann, assumes a linear approach, stepping commands. And it seems that such an approach close to the limit of their own abilities, because computer engineers are trying to do more than the latest "non-von Neumann" computer architecture in the MDM processes are not organized serially and in parallel — as in our brain.

Ah well describes the advantages of these processes, one of the founders of the report IBM Dharmendra EHPM: "functionality, low power consumption, small footprint," "human brain possesses the awesome power — the ability to receive, comprehend, learn, act, and lead the interaction, and for all that consumes less power than an electronic light bulb, and it takes less volume than the two-liter bottle of soda. " Here's a video from IBM "era of cognitive systems" where possible to see how the chip TrueNorth plays in the elementary computer game Pong — "just like a man."

In this modern computer architecture broken calculations (for their answer "neurons"), memory ("synapses"), information exchange ("axons" and "dendrites") — similar to the device of the brain. Just like real, computer "neurons" receive information from adjacent thousand "neurons" treated and its transmitted signals, and computer "synapses" bind "neurons" and hoarding information.

In another work, the next time in this area, researchers from Germany, the euro project participants BrainScales, did Spikey c chip analog components, virtually reproducing the work of neurons and synapses that simulate their electronic behavior. In NewScientist one of the founders of Karlheinz Meyer explains, just as in real brain neuron sends a signal when the voltage at the electronic membrane achieves a certain level, and the chip capacitor plays the role of a neuron becomes conductive when a specific cell voltages. Analog components that act as synapses possess different levels of resistance — in exactly the same connection between neurons become stronger or weaker depending on their frequency of use.

This line of development of computer technology is not the only thing there is, for example, the idea of a quantum computer, which, by the way, are related works of the last winners of the Nobel Prize in Physics. But specifically, research in the field of computer simulation of the brain entice non-specific media attention proximity to own speculative ideas of the creation of artificial intelligence.

Although researchers at IBM have identified that their project was not about the simulation of brain activity in biological sense, in comments to their work enthusiasts said that is a fundamental step towards the creation of a true artificial brain, "according to all estimates, we need many more- many years to recreate the true behavior of the human brain, the most complex object in the known universe, but, at least, it seems there has been progress. " Skeptics remind: human brain investigated so weak that it is not yet clear how, in fact, to build his likeness.

For this personal discussion mountain towering ancient debate about whether it is possible for a computer, an algorithmic way to get out. Determinism versus free will. One approach illustrates a demon — Laplace thought experiment French mathematician XVIII-XIX centuries, according to which, if a powerful mind will know the position of all the bodies in the universe (from the largest to the smallest bodies of atoms) and all the forces acting on them, he learns of a single formula and will build the future is as clear how the past. On the other hand, our contemporary, physicist Roger Penrose suggests quantum model of brain activity, in which the processes of thinking fundamentally non-computable.

Penrose such makarom destroys building a new determinism which, seems to imply the creation of artificial intelligence mechanistic means — gathering brain neuron by neuron. In general, even those who consider possible ways to model the human consciousness of cybernetics, the vision of this process is more complex. Here, for example, a lecture Misha Burceva to what Artificial intelligence is the result of studying the evolution of artificial neural networks controlling computers or robots (and artificial neural City — it's just that, working on what IBM and other researchers). The lecture is put a nice question — it will be the evolution of man, which extends the phenotype own artificial intelligence, or it will be the emergence of a new kind — artificial intelligence? (Here, though what will come in the memory of the discussion between the rough and suave Korneev Amperyanom: "How do we know that the purpose of nature — do Amperyana friend? Maybe the purpose of nature — is the creation of undead hands Amperyana comrade").

Illustrates the complexity of the tasks next November's study — a group of physicists published a paper on the similarity of the evolution of complex systems — the brain, the web and the universe. One of the founders of Dmitry Kryukov, in an interview with Radio Liberty, described the process of creating a graph of all the simple events in the history of the universe, and it sounded so hard and great that even the kind can not be brought within the framework of today's article — in the coming days are absolutely we will publish the interview. According Kryukov, it became clear that the dynamics of the universe turned out to be similar to the growth dynamics of very different systems, such as web, social networks and brain"Explanation of this marvelous coincidence of today we do not have, but most likely it is that there are some fundamental laws of the unknown, of which the individual case, the Einstein equations describing the dynamics of the universe in general relativity, and unknown to the equations of motion web development, social networking, and the brain in common, we also today do not know. " Hooks emphasizes that deals
with the empirical, observable dynamics, and we do not know the equations of motion for all of these complex systems, but if you knew, would herald their development. What is a demon of Laplace?

Similar size and complexity of tasks creation of an artificial brain can be profitable and public land, taking into account, they view the consequences of this scientific breakthrough, even his followers. The emergence of super-human intelligence described by the term "technological singularity" in a famous article in 1993 mathematician and fiction writer Vernor Vinge. After a point of singularity, he believed, the human era should end in a rapidly: once in response to a question about whether the computers are ever alike as intelligent as a human, Vinge said: "Yes, only a very short time", meaning that cost computers to catch up with the man, they will not stop and will go forward simultaneously. In the article the era's population was driven more time in the range of 30 years — in other words, until 2023.

So far, but can not see where it might seem artificial mind, and there is no fear that the chain of transistors and wires stir, Husky says "hello", joking pass the Turing test (designed to distinguish human from a machine) and leave for Christmas shopping.

So it is absolutely calm — the last message in this article about the study: a group of scientists from the Cambridge Institute organized a research center of existential risks, and which carried the risk of the artificiality of the mind.

Sleep quietly. Scientists will look after the robots run amok.

Like this post? Please share to your friends: