Researchers from George Washington University in the US discovered that using photons within neural network (Tensor) processing units (TPUs) could create more powerful and power-efficient AI. They have achieved a breakthrough in the development of artificial intelligence by using light instead of electricity.
Tensors are a type of data structure used in linear algebra, and like vectors and matrices, you can calculate arithmetic operations with tensors. Many of the operations that can be performed with scalars, vectors, and matrices can be reformulated to be performed with tensors.
The new approach significantly improves both the speed and efficiency of machine learning neural networks to replicate the work of a human brain in order to teach itself a task without supervision. Current processors used for machine learning are limited in performing complex operations by the power required to process the data. The more intelligent the task, the more complex the data, and therefore the greater the power demands.
Conventional machine learning neural networks are also limited by the slow transmission of electronic data between the processor and the memory.
The team of scientists revealed that their photon-based TPU was able to perform between 2-3 orders of magnitude higher than an electric TPU. Mario Miscuglio, a member of the team said,
We found that integrated photonic platforms that integrate efficient optical memory can obtain the same operations as a tensor processing unit, but they consume a fraction of the power and have higher throughput. When opportunely trained, can be used for performing interference at the speed of light. Photonic specialised processors can save a tremendous amount of energy, improve response time and reduce data centre traffic.
There are lots of uses of this new TPUs. Upcoming 5G and 6G networks can use these lightning fast processors to decrease latency. A well as the data centres with vast amounts of data processing could also get huge benefit from these. Famous machine learning libraries like Google's TensorFlow and Amazon's SageMaker could be more efficient by using this new technology.