A Silicon Valley startup has developed an optical processor capable of increasing the performance and efficiency of AI systems.
The Artificial Intelligence is advancing at such a speed that the industry is demanding more and more innovations to accelerate IT processes while keeping power consumption to a minimum.
Conventional AI processors are not best suited to perform those tasks that require a very high computing power. These constraints lead to delays which, for example, hinder the operation of search engines using machine learning or the guide algorithm of autonomous vehicles or drones.
And in this context, the startup Luminous Computing wants to offer a solution to the excess of hardware and boost the performance of semiconductor chips. The answer to this bottleneck is to use lasers for emitting light through small structures in the chips. Optical processors fit perfectly as a way to manage a large number of operations that run Artificial Intelligence systems.
These little structures function as light wave guides, thus achieving a small loss of energy. The startup aims to use different colors of light to transfer bits via simultaneous waves. The process can perfect the operation of the electronic chips and make them more efficient data transport than conventional chips.
These new optical processors are best suited for Artificial Intelligence methods that require a high level of computing. The differential element is that they manage to transport a large amount of data from simultaneously and they need a lot less power.
Luminous Computing is supported by figures of great relevance in the tech industry, like Bill Gates. The current prototype of the optical processor is designed to offer a three times higher energy efficiency like conventional AI chips. This increase in the pace of operations can contribute to achieving a unprecedented speed.
If you want to continue reading about Artificial Intelligence and virtual assistants do not miss this post about Aura, the Artificial Intelligence of Telefónica.