- Optical computing uses light instead of electricity to process complex data.
- The digital twin eliminates long waits for shared optical hardware.
- Virtual optical systems mirrored real hardware with remarkable accuracy.
Optical computing has emerged as a promising alternative to traditional electronic systems struggling with increasingly large-scale deep learning and artificial intelligence workloads.
By taking advantage of the physical properties of light, including interference and diffraction, optical computing systems offer faster speeds, better energy efficiency, and more robust parallel processing capabilities.
Chinese researchers have now proposed a digital twin model that fundamentally changes the way these complex systems are developed and tested.
Why physical hardware became a bottleneck for researchers
Traditional optical computing systems face a persistent challenge, as task development relies heavily on direct access to physical hardware platforms.
When multiple researchers need to work with the same system, they typically wait in line, then repeatedly adjust parameters and perform error calibration before any genuine calculations can begin.
Once one user is done, the next must often reset the entire system state, making parallel research between competing projects nearly impossible.
That cycle of waiting, adjusting, and recalibrating increases trial-and-error costs while severely limiting overall research efficiency.
To address that bottleneck, the researchers developed what they call the Twin Digital Optical Computing System, or DT-OCS, published in Opto-Electronic Advances.
The framework builds a digital model that reproduces the input-output responses of a physical optical computing system through different configuration parameters entirely within software.
If the physical system resembles an expensive and busy real machine, the researchers describe DT-OCS as functioning as a high-fidelity simulator running alongside it.
Testing image classification within a virtual twin before touching real hardware
Using a high-speed optical computing system paired with a silicon photonic feature computing chip, the research team tested DT-OCS in image classification and sequential decision-making tasks.
The results showed that the configuration parameters trained and optimized within the digital twin were transferred directly to the physical system without requiring additional adjustments.
Performance of the task on the physical hardware closely matched the predictions of the digital model, validating both the fidelity and transferability of the entire approach.
Because training and optimization primarily occur within the digital domain, researchers can now perform multiple different tasks simultaneously instead of queuing for shared access to hardware.
The team has also made the DT-OCS framework and its associated data sets openly available.
This will allow other researchers to perform training and validation without even touching the physical equipment.
According to the researchers, they designed DT-OCS as “a reproducible, accessible, and scalable software resource for broader sharing and validation.”
Openness effectively transforms optical computing from a specialized resource limited by device availability to something closer to a reproducible and shareable research platform.
The researchers argue that future optical computing systems should combine physical hardware with openly available digital models that offer equivalent computational behavior.
By comparing how modern transportation relies on both physical roads and continually updated digital maps, they suggest that mature optical computing platforms need a similar dual structure in the future.
Via EurekAlert
Follow TechRadar on Google News and add us as a preferred source to receive news, reviews and opinions from our experts in your feeds.
