A silicon-photonic (SiPh) computer chip that runs on light instead of electricity has been developed by a team of researchers at the University of Pennsylvania, according to Interesting Engineering.
This invention has the potential to improve data transfer speed and lower electricity usage, which will be especially helpful for training artificial intelligence (AI) models.
Under the direction of Professor Nader Enghata, the group created the SiPh chip, which uses light to perform mathematical calculations by taking use of its unmatched data transfer speed.
Scalability is made possible by using silicon that is widely available. The chip’s primary role is to conduct vector-matrix multiplications, a crucial mathematical operation for the construction and operation of neural networks, which is essential for modern artificial intelligence models.
Most importantly, the researchers solved issues with data privacy related to parallel computing. A layer of protection against possible hackers is added because the SiPh semiconductor can do numerous computations at once, negating the need to keep data in working memory.
With possible uses ranging from replacing graphics processing units (GPUs) in AI model training and classification, the SiPh chip is ready for rapid adoption.
According to University associate professor Firooz Aflatouni, the SiPh platform may be easily integrated into the current AI infrastructure, resulting in faster computations and less electricity usage, which would represent a major advancement in computing technology.
I am a dedicated student currently in my seventh semester, pursuing a degree in International Relations. Alongside my academic pursuits, I am actively engaged in the professional field as a content writer at the Rangeinn website.