IBM researchers reverse engineered a macaque's brain to begin the process of building its own electronic brain.

Engineers at IBM Research Labs are not content to create a genius robot that can defeat any human opponent in the Jeopardy quiz, not creating a comparable human brain. New smart machines are not available.

The show's defending champion Watson is undoubtedly smart, but it is still a recognized computer. And the new inventions of IBM engineers will be completely different things - IBM is trying to build an electronic brain from scratch.

This new field, called cognitive computing, takes the concept of computers to a whole new level. Earlier this week, Dharmendra Modha of the IBM Almaden Research Center introduced to a house of scientists what cognitive computing can do and how IBM will proceed. Invented this machine that is the same way we think.

Modha first described the challenges it faced, including neuroscience, supercomputing, and nanotechnology.

The human brain integrates memory and storage into one, weighs less than 3 pounds and takes up about two liters, but is more energy efficient than a light bulb. It runs like a massively parallel distributed processor. It is event driven, that is, it reacts to things in its environment. Less energy is consumed in the active state and less in the rest state. It is a reconfigurable, fault-tolerant learning mechanism. It is very good at pattern recognition and relationship combing.

On the other hand, the storage and processing of computers is separate. Most of its operation is sequential in order and controlled by a clock. This clock is like a conductor of a military band, driving every instruction and every piece of data to the next position - just like having enough space to grab a chair game. As clock speeds drive data processing faster, power consumption increases dramatically, and these machines require a lot of power even during sleep. More importantly, programming is essential. They are connected by wires and are prone to failure. They are good at executing predefined algorithms and analysis work.

Using a $41 million grant from the Defense Advanced Research Projects Agency, scientists at the Almaden Laboratory are in a neuromorphic AdapTIve Plas TIc Scalable Electronics; Under the abbreviation of SyNAPSE "Synap", the project began to construct the brain.

A nearly similar feature between the brain and the computer is the setting of roles for cell types - computer components such as neurons, axons, and synapse-compatible processors, communication links, and memory. This match is not very accurate, as the division of labor between all brain cells is not as clear as the computer components. But the key is that the brain cells are very close to each other, and the activity in any particular brain tissue is caused by the activity of neighboring tissues. In other words, one idea stimulates another idea.

Modha and his team set out to map the brain's wiring diagram and start synthesizing, which is a daunting task because the brain has 22 billion neurons and 220 trillion synapses. In May 2009, a team successfully simulated the creation of a system of one billion neurons, roughly equivalent to the brain of a lower mammal. But its processing speed is one-thousandth of the speed of actual events, not enough to perform the tasks that Maud calls a generalization of "4F": food, fight, flight, and mating (maTIng) ).

But the structure of this machine is completely different from today's business computers. The memory and processor components are tightly coupled. It has no clocks, operations are asynchronous, event-driven, that is, they do not have a predetermined order or schedule. And they rely on learning rather than programming. Just like humans.

Part of the practice of reducing energy consumption to a level similar to the human brain is to not store temporary results (in industry terms, it is not "cached"). Perception can stimulate action, and action is perceived and stimulates further action. And so on.

The team recently simulated a smaller version of the hardware built into the brain, with only 256 neurons, 262,000 programmable synapses, and 65,000 learning synapses. Fortunately, the energy consumption of this machine is on the same order of magnitude as the energy consumption of the brain. With its original ecological ability, this "small human brain" can perform space navigation, machine vision, pattern recognition and associative memory, and can generate hypotheses based on evidence. It has an "eye of the mind" that identifies the pattern behind the thing, for example, based on a scribbled note, making a more accurate guess of the number it actually represents. It is already better than our precambrian ancestors.

Modha pointed out that this type of reasoning is very similar to the function of the typical right hemisphere of the brain: intuitive, analogous and integrated. Not satisfied with half of the brain, Modha envisions adding a typical von Neumann computer to perform the left hemisphere's reasoning behavior, allowing the two hemispheres to share information, just like a real human brain.

When it goes on sale, I want to let my own brain go on vacation, and give the task of thinking to this masterpiece of Modha.

Oh, by the way, if you want to know if the "synaptic" project can already make Watson retired, the answer is - this is far from impossible. Watson is still alive and well, and is continuing to invest in new, more practical uses.

For example, because the entrants of the Crisis Edge game don't have the option of “call for help”, Watson is limited by the data directly loaded (not connected), but in the latest application area of ​​“Watson” technology – medical Diagnostics - The Internet was easily added to its corpus, allowing Watson to search a wider range of unstructured data before providing answers.

In the competition, Watson had to press the answer button faster than its human players, but doctors who ask for advice on a strange set of symptoms can be more willing to wait for half an hour or more. Therefore, Watson can make a more prudent decision. At work, Watson is a serious robot.

Mylar Speaker

The Mylar Speaker solutions come in a variety of shapes and sizes. We can provide various mounting configurations and performance alterations to fit any industrial application. Our Mylar speakers produce excellent sound output (dB) at specified frequency ranges. The unique characteristics of a Mylar cone allow us to keep tight tolerances during the manufacturing process. Additionally, Mylar is easily and consistently moldable, creating a cost-effective solution, time after time. It`s also beneficial in applications that are exposed to excessive moisture or humidity since Mylar has a high resistance to environmental factors.

Mylar Speaker,Mylar Tweeter,Mylar Cone Speaker,Cellphone Mylar Speaker

Jiangsu Huawha Electronices Co.,Ltd , https://www.hnbuzzer.com

Posted on