The new chip opens the door to artificial intelligence computing at the speed of light

Engineers at the University of Pennsylvania have developed a new chip that uses light waves instead of electricity to perform complex mathematical operations necessary for training artificial intelligence. This chip has the potential to fundamentally accelerate the processing speed of computers while reducing their energy consumption.

The design of a silicon photonic chip was the first to combine the Benjamin Franklin Medal winner with H Professor Nedwell Ramsey Nader Engheta's pioneering research on manipulating materials at the nanoscale to use light for mathematical calculations is combined with the SiPh platform, which uses silicon as a cheap and abundant element for large-scale production of computer chips.

The interaction between light waves and matter represents a possible way to develop computers that have replaced the limitations of today's chips, which are basically based on the same principles as chips in the early stages of the computing revolution in the 1960s.

In a paper published in Nature Photonics, Engheta's team, along with the team of Associate Professor of Electrical and Systems Engineering Firoz Aflatouni, described the development of a new chip.

"We have decided to work together," Engheta said, leveraging the fact that Aflatouni's research team has pioneered nanoscale silicon devices.

Their goal is to develop a platform to perform so-called vector matrix multiplication, which is the core mathematical operation of neural network development and functionality. Neural networks are the computer architecture of today's artificial intelligence tools.

Engheta explained, "You're not using highly uniform silicon wafers, but making the silicon thinner, such as 150 nanometers," but only in specific areas. These height changes - without adding any other materials - provide a way to control the propagation of light through the chip, as the height changes can be distributed to cause light to scatter in specific modes, allowing the chip to perform mathematical calculations at the speed of light.

Aflatouni said that due to restrictions imposed by commercial foundries producing chips, this design is ready for commercial applications and may be applicable to graphics processing units. With the widespread interest in developing new artificial intelligence systems, the demand for graphics processing units has surged.

"They can use silicon photonics platforms as additional components," Aflatouni said, "and then you can accelerate training and classification speed.".

In addition to faster speeds and lower energy consumption, Engheta and Aflatouni chips also have privacy advantages: because many calculations can be performed simultaneously, sensitive information does not need to be stored in the computer's working memory, making future computers driven by this technology almost impossible to crack.
"No one can invade non-existent memory to access your information," said Aflatouni.

Other co authors include Vahid Nikkhah, Ali Pirmoradi, Farshid Ashtiani, and Brian Edwards from the School of Engineering at the University of Pennsylvania.

Source: Laser Net