The contribution This AI uses light instead of electronics – and is incredibly quick by Felix Baumann first appeared on Basic Thinking. You always stay up to date with our newsletter.
Photonic neural networks improve monitoring through light instead of electricity. This enables the analysis of large sensor data in real time.
Scientists from the University of Nanjing have developed a new AI systemthe data not processed with electricity, but with light. It not only works extremely quickly, but also energy -efficient. The technology, a so-called photonic neuronal network, was successfully combined for the first time with a DAS system (Distributed Acoustic Sensing).
This analyzes fiber optic cables to recognize fine vibrations – for example for earthquake warnings, the monitoring of railway lines or lower sailing cables. The problem: the resulting amounts of data are huge and can hardly be processed in real time with classic electronics.
How photonic neural networks change monitoring
However, photonic neural networks only process signals optically. This means that data processing is evaluated with light instead of electrical currents. The heart of the technology is the Time-Wavelength Multiplexed Photonic Neural Network Accelerator (TWM-PNNA).
This uses various laser wavelengths to carry out mathematical operations – so -called folds. This allows sensor data to be analyzed at lightning speed. Photonic neuronal networks significantly improve monitoring, especially if large amounts of data must be evaluated reliably and without delay.
For this, the researchers had to master technical challenges such as frequency shifts in light modulation that could affect the accuracy of the results. Through methods such as push-pull modulation, the system achieved a recognition accuracy of over 90 percent.
This is almost at the level of conventional electronic systems. At the same time, energy consumption is significantly lower: the new approach creates up to 21 trillion computing operations per watt – a multiple of what current GPUs can do.
Potential for infrastructure monitoring
Another advantage is the efficiency of the structure. With the help of so-called Pruning techniques, the photonic networks can be severely reduced without accepting noticeable performance losses. This not only makes the technology powerful, but also cheaper and easier in production.
In the long term, photonic neuronal networks offer great potential for infrastructure analysis. Due to their combination with sensor networks, train tracks, pipelines or seismic regions can be monitored quickly, precisely and with minimal energy use.
Especially in situations in which real -time processing is decisive, photonic neuronal networks significantly improve surveillance. You could become an important part of modern early warning systems and digital infrastructure control.
Also interesting:
- New charging system: BYD wants to charge electric cars in 5 minutes
- New cyber security test for real-time systems (RTOS)
- Porn filter for operating systems: federal states tighten youth protection
- Solar system transforms salt water into drinking water – without batteries
The contribution This AI uses light instead of electronics – and is incredibly quick by Felix Baumann first appeared on Basic Thinking. Follow us too Google News and Flipboard.
As a Tech Industry expert, I am fascinated by the potential of this new AI technology that utilizes light instead of traditional electronics. The use of light for computing offers a promising alternative to current methods, potentially leading to faster processing speeds and more energy-efficient systems.
The speed at which this AI operates is particularly impressive, showcasing the power of utilizing light for data processing. The ability to harness light for computing could revolutionize the way we approach AI technology, opening up new possibilities for applications in various industries.
I am excited to see how this technology develops and how it could potentially shape the future of AI and computing as a whole. The possibilities for innovation and advancement in this field are truly exciting, and I look forward to seeing how this technology will continue to evolve and improve in the coming years.
Credits