Fabrican chips de computadora similares al cerebro… ¡con miel!

El futuro de la computación podría estar en la miel. Como suena. Un equipo internacional de científicos ha comprobado que el fluido que elaboran las bejas could be a ‘sweet solution’ to develop ecological components for neuromorphic computerssystems designed to imitate neurons and synapses of the human brain, y que son much more fast and use much less energy than traditional computers.

A team of engineers from Washington State University (WSU) has demonstrated a new way to make these systems also “more organic”. En un estudio publicado en ‘Journal of Physics D’, los inquisitors show that honey can be used to construct a ‘memristor’, a component similar to a transistor that can not only process but also store data in memory.

“Este es un dispositivo muy pequeño con una structure simple, pero tiene functionalities very similar to those of a human neuron“, pointed out Feng Zhao, co-author of the studio.

“This means that if we can integrate millions or thousands of millions of honey memristors together, then we can convert them into a neuromorphic system that funcionará de manera muy similar a un cerebro humano“, he adds.

Zhao and the principal author of the investigation, Brandon Sueoka, created for the study memristors to process the miel in a solid form and place it between two metal electrodes, forming a structure similar to human synapsis.

Then they tested the capacity of memristors for honey Imitar el trabajo de las synapsis with high speeds of 100 and 500 nanoseconds on and off, respectively.

‘Memristores’ del tamaño de un cabello

The memristors also emulated the functions of synapses known as ‘peak time dependent plasticity’ and ‘peak speed dependent plasticity’, which are responsible for learning processes in the human brain and retaining new information in neurons.

The engineers created honey memristors of the approximate size of a human hair. The team directed by Zhao now plans to develop them on a nanoscale, approximately a thousand times smaller than a human hair, and group muchos millions or even thousands of millions for crear un sistema informatics neuromórfico completo.

Dashboard. pixabay

Los sistemas informáticos conventionales se basan en lo que se denomina arquitectura de von Neumann, que implica una entrada, por lo general desde un teclado y un ratón, y una salida, como puede ser el monitor. It also has a CPU, the central processing unit, and RAM, the storage memory.

La data transfer through all these mechanisms, from the entrance to the processing, from the memory to the exit, requires a lot of energy, at least compared to the human brain, explains Zhao.

Put an example: la supercomputer Fugakudeveloped by Fujitsu for the Computer Sciences Center RIKEN in Kobe (Japan), usa más de 28 megawatts (the power installed in a medio-sized wind park, like the one in Valdefuentes, in Huelva) to function, while el cerebro usa solo alrededor de 10 a 20 watts.

The human brain has more than 100,000 million neurons with more than 1,000 billion synapses or connections between them. Each neuron can process and store data, which makes the brain much more efficient than a traditional computer, and the developers of neuromorphic computer systems pretend to imitate that structure.

Biodegradable and renewable solutions

Various companies, including Intel and IBM, have been launched chips neuromorficos que tienen the equivalent of more than 100 million “neurons” per chip, but this is not yet known about the number existing in the brain. Un elevado número de desarrolladores también sigue utilizando los mismos materiales no renovables y toxicos que se utilização actualmente en los chips de computadora conventionales.

Many researchers, including Zhao’s team, are there buscando solutions biodegradable and renewable para usar en este nuevo y prometedor tipo de computación.

Un apicultor comprueba una de las colmenas. unsplash

Zhao is also leading research on it uso de proteins y otros sugarescomo los que se encuentra en las hojas de aloe vera, pero ve un major potencial en el fluido que elaboran las bejas.

“La miel no se echa a perder”, destaca. “Tiene una concentración de humidad muy baja, por lo que las bacteria no pueden sobrevivir en él. Esto significa que These computer chips will be very stable y confiables durante mucho tiempo”, he added.

The Honey Memristor chips developed at WSU should tolerate lower levels of heat generated by neuromorphic systems than no se calientan tanto como las computadoras tradicionales. Los memristores de miel también reduce electronic wasteauguran los inquisitors.

“When we want to destroy devices that use computer chips made of honey, we can dissolve them easily in water,” explains Zhao. “Debido a estas propiedades especiales, la miel es muy útil para crear sistemas neuromórficos renovables y biodegradables“.

“Esto también significa que al igual que las computadoras convenciales, los usuarios must avoid spilling coffee on them,” concludes Zhao with grace.

Informe de referencia: https://iopscience.iop.org/article/10.1088/1361-6463/ac585b


Contact of the Medio Ambiente section: crisisclimatica@prensaiberica.es


Leave a Comment

Your email address will not be published. Required fields are marked *