Great news to Artificial Intelligence developers!! Now, the processing of enormous amount of neural data is just a matter of seconds. An international research team led by the Swinburne University of Technology in collaboration with Monash and RMIT universities has come with the fastest and powerful optical neuromorphic processor for AI. 10 trillion operations within a second and processing of huge amount of data makes it powerful. This is truly an enormous breakthrough in the AI industry. We know AI finds useful for different applications like facial recognition, healthcare, speech therapy, communication, and other areas. So, the need for an optical neuromorphic processor is demanding.
The Road to Optical Neuromorphic Processor
David Moss (Swinburne's Processor), Dr. Xingyuan Xu (Professor Monash University), and Arnan Mitchell (Professor of RMIT University) are the team that led to this discovery. They demonstrated that the processor is 1000 times faster and it can process ultra large scale images. This feature is one of the best achievements of the AI industry. This achievement was possible due to a technique using optical micro-combs.
Till yesterday, a large number of parallel processors were required to operate beyond 100 TeraOP/s. But now with the Optical neuromorphic processor which uses a different technique to interleave the data, spatial dimensions, and wavelength through an optical micro-comb; everything is in seconds.
Optical Micro-combs are devices made of 100's of infrared lasers. These lasers are of high quality and act like a rainbow. Compared to an optical source, micro-combs are small, fast, lighter, and cheap.
Swinburne's Professor Moss Says that "integration of micro comb chips has enabled in the advancement of communication, processing and information sector."
"The processor can act as a front end with high bandwidth for any electronic or optical-based neuromorphic hardware," says Dr. Xu (Swinburne, Monash University).
Dr. Xu, also explains "How the processor would look like. Through the efficient use of optical micro-combs, it shows us how we scale the power of the processors"
MIT's Professor Mitchell says that "Optical Neuromorphic processor will provide a huge impact in the processing and communication sector."
Advantages of the Optical Neuromorphic Processor
- Ultrahigh bandwidth
- Massive information processing
- Reduce cost
- Low Energy Consumption.
- High Speed
- High Processing Power
- Better accuracy.
In a conclusion, the increasing demand for AI insists on the need for a processor with the above advantages. An optical-based neuromorphic processor is all the best.