Nano brain

A nano brain is a conceptual device with massively parallel computational abilities, following the information processing principles of the human brain. This machine assembly would serve as an intelligent decision making unit for nanorobots.
Necessity
Computing in the 20th century was confined inside a box, or machine, called computer or supercomputer; now, several parameters around us compute (Internet of Things). Earlier, we used to have small amounts of information stored in a book or server, and used to upload and download as required. In this and the next century, this could reverse, since the locations where we store information are becoming astronomically large. Without extending computing beyond serial logic, we could get isolated in the information domain without getting connected to the desired point. Human brain follows pattern-based computing like chaos, cellular automaton wherein millions of pixels of a particular image is processed at a time. The mechanism appears extremely slow and inaccurate. However, as the complexity of information increases, it performs more credibly than manmade supercomputers. A human brain can read captcha letters in seconds; a supercomputer can not do that in a finite time. Exponential increment of information generates a serious challenge for command, control and processing when connectivity among these information also increases exponentially. Since software uses a sequential approach to analyze connectivity, to shrink infinite complexity into a finite limit, mechanism of processing infinite information has to be embedded inside the hardware. Nano brain is such a device that physically addresses nearly infinite possible connections in seconds, alleviating the singularity in the software. This concept has the potential to solve at least three bottlenecks of human civilization, providing necessary intelligence to the robots, executing jobs without conventional power supply and finally, resolving the many-body problems which are in abundance in nature.
Conceptual novelty of the hardware
Historically, equivalent circuits have been proposed for neurons and even for the central nervous system. Creating an equivalent circuit is a reliable means to understand the electronics of a complex device as it defines the device in terms of fundamental circuit elements. The functional principle of a nano brain architecture is to exhibit "one-to-many communication at a time" among the constituent decision making units. By conventional circuit theory, it is parallel circuiting of elements. Since the conformation of wiring path changes along with electronic charge transport in the circuit, equivalent circuit would change continuously. The possible combination of such circuits is astronomically large therefore instead of defining a function for an evolving equivalent circuit, the concept of cellular automaton has been introduced. In addition, due to spherical design, information spreads out from the center of the sphere and again reflects back to the center from the outer surface. Every single atom in the spherical nano-brain experiences a continuous interference of feed forward information wave. Thus, the concept of circuit is violated here as collective evolution of a potential distribution in a 3D space at a time can not be represented as a linear sequence of events in discrete times.
Evolving hardware
Biological neural networks in the human brain evolves continuously throughout the lifespan, allowing the brain to form folds. There are several attempts to realize evolutionary circuits; however, the majority of these attempts assemble a few static circuits and choose one of them during computation. The human brain's evolution is functionally different: neurons change connections to make short-cuts. These routes lead to faster decision making (referred to as increasing efficiency through learning). A nano-brain changes connections between different sub-processors in a very similar fashion; therefore, it learns with experience. Since no hardware restriction is imposed in the nano-brain, the possibilities of changing are enormously large (not astronomical since restriction due to resource limitation imposes an upper limit); however, that number ranges in the order of millions compared to tens in the present evolving hardwares.
Multilayered decision making
There are several layers of subprocessors one top of another that constitute the nano brain, the bottom-most layer connects to the external machines or sensors and the top-most layer carry the fundamental rules that are never changed during nano brain computation. If nano brain is made of cellular automaton then number of cells decreases in every layer as computation transits upward. The embedded cellular automaton cluster that represent entire nano brain, follows two different classes of cellular automaton rules. First class of rules are those which are followed in the cellular automaton grid, and the other class of rules are basically the transition rules between two cellular automaton layers, each layers are termed as sub-processors.

Main Menu

See Also

Main Menu