“Unhackable” Computer Chip Works on Light

This post is also available in: עברית (Hebrew)

Researchers from the University of Pennsylvania developed a new computer chip that uses light instead of electricity, which can improve AI model training by improving the speed of data transfer and reducing the amount of electricity consumed.

While humanity’s computers are getting increasingly faster and are processing more and more information, computing technology is still working on the principles that were first used in the 1960s. one of the ways researchers have been working on improving computing power is by developing computers that are quantum-based, but these computers are years from becoming widely. At the same time, the recent explosion of AI models in technology resulted in a demand for computers that can process large sets of information. All this means an increase in demand for processing power with inefficient computing systems, which results in high energy consumption.

According to Interesting Engineering, a team from the University of Pennsylvania led by Nader Enghata, a professor at the School of Engineering and Applied Science, designed a silicon-photonic (SiPh) chip that can perform mathematical computations using light, which is the fastest means of transferring data known to humanity. In addition, the use of widely abundant silicon ensures the technology can be scaled quickly.

The researchers aimed to design a chip that can perform the mathematical computation needed to develop and operate the neural networks that are critical when developing architecture to power AI models. Since the chip is made with silicon, the researchers could have reinvented the fabrication process completely. Instead, they reduced the height of the chip in specific regions to control how light propagates inside – by controlling the height variations, the research team ensured that the light inside the chip traveled only in a straight line.

Associate professor in Electrical and Systems Engineering at the University Firooz Aflatouni said that the chips could replace the GPUs that companies use to train and classify their AI models, and recommended that the SiPh platform serve as an add-on to existing infrastructure being used by AI companies.

The SiPh chips can not only perform computations faster and with less electricity consumption, it can also address data privacy concerns. Since it can perform multiple computations in parallel, there is no need to store the information in a working memory while the computations are performed, and so no one can hack into a non-existing memory to access your information.