New reservoir computer marks first-ever microelectromechanical neural network application

October 16, 2018, American Institute of Physics
A single silicon beam (red), along with its drive (yellow) and readout (green and blue) electrodes, implements a MEMS capable of nontrivial computations. Credit: Guillaume Dion

As artificial intelligence has become increasingly sophisticated, it has inspired renewed efforts to develop computers whose physical architecture mimics the human brain. One approach, called reservoir computing, allows hardware devices to achieve the higher-dimension calculations required by emerging artificial intelligence. One new device highlights the potential of extremely small mechanical systems to achieve these calculations.

A group of researchers at the Université de Sherbrooke in Québec, Canada, reports the construction of the first built with a microelectromechanical system (MEMS). Published in the Journal of Applied Physics, the exploits the nonlinear dynamics of a microscale silicon beam to perform its calculations. The group's work looks to create devices that can act simultaneously as a sensor and a computer using a fraction of the energy a normal would use.

The article appears in a special topic section of the journal devoted to "New Physics and Materials for Neuromorphic Computation," which highlights new developments in physical and materials science research that hold promise for developing the very large-scale, integrated "neuromorphic" systems of tomorrow that will carry computation beyond the limitations of current semiconductors today.

"These kinds of calculations are normally only done in software, and computers can be inefficient," said Guillaume Dion, an author on the paper. "Many of the sensors today are built with MEMS, so devices like ours would be ideal technology to blur the boundary between sensors and computers."

The device relies on the of how the silicon beam, at widths 20 times thinner than a human hair, oscillates in space. The results from this oscillation are used to construct a virtual neural network that projects the input signal into the higher dimensional space required for neural network computing.

In demonstrations, the system was able to switch between different common benchmark tasks for neural networks with relative ease, Dion said, including classifying spoken sounds and processing binary patterns with accuracies of 78.2 percent and 99.9 percent respectively.

"This tiny beam of silicon can do very different tasks," said Julien Sylvestre, another author on the paper. "It's surprisingly easy to adjust it to make it perform well at recognizing words."

Sylvestre said he and his colleagues are looking to explore increasingly complicated computations using the silicon beam device, with the hopes of developing small and energy-efficient sensors and robot controllers.

Explore further: A new brain-inspired architecture could improve how computers handle data and advance AI

More information: "Reservoir computing with a single delay-coupled non-linear mechanical oscillator," Journal of Applied Physics (2018). DOI: 10.1063/1.5038038

Related Stories

Chip ramps up artificial intelligence systems' performance

September 24, 2018

Princeton researchers, in collaboration with Analog Devices Inc., have fabricated a chip that markedly boosts the performance and efficiency of neural networks—computer algorithms modeled on the workings of the human brain.

Brain function partly replicated by nanomaterials

July 12, 2018

The brain requires surprisingly little energy to adapt to the environment to learn, make ambiguous recognitions, have high recognition ability and intelligence, and perform complex information processing.

Recommended for you

Physicists build fractal shape out of electrons

November 12, 2018

In physics, it is well-known that electrons behave very differently in three dimensions, two dimensions or one dimension. These behaviours give rise to different possibilities for technological applications and electronic ...

Atomic parity violation research reaches new milestone

November 12, 2018

A reflection always reproduces objects as a complete mirror image, rather than just its individual parts or individual parts in a completely different orientation. It's all or nothing, the mirror can't reflect just a little. ...

Innovative experimental scheme can create mirror molecules

November 12, 2018

Exploring the mystery of molecular handedness in nature, scientists have proposed a new experimental scheme to create custom-made mirror molecules for analysis. The technique can make ordinary molecules spin so fast that ...

0 comments

Please sign in to add a comment. Registration is free, and takes less than a minute. Read more

Click here to reset your password.
Sign in to get notified via email when new comments are made.