- High-Power Lasers and Ethanol create created holograms to cod data
- AI was trained to decode the data with 90-95% accuracy
- The optical coding technology will not get on the market quickly, while some shortcomings hinder its usability
A team of Greek scientists has developed an optical coding system that could possibly make traditional hacking methods superfluous.
In a research paper published by OpticalThe researchers behind the system revealed that it combines artificial intelligence (AI) and laser -generated holograms to offer data protection at a high level that could be virtually impossible to crack.
In contrast to traditional coding methods, which depend on mathematical algorithms, this optical approach uses the physical properties of light, making it resistant to attacks of even the most advanced computers, including quantum machines.
How it works: Strambling Light for Security
“From rapidly evolving digital currencies to governance, health care, communication and social networks, the demand for robust security systems to combat digital fraud continues to grow,” said Stelios Tzortzakis research team, of the Institute of Electronic Structure and Laser, Foundation for Research and Laser, Foundas, Foundas, Research and Laser, Foundas, Research and Laser, Foundas and Laser Technology, Research and Laser, Foundas and Laser, Foundas and Laser, Foundas and Laser, Foundas and Laser, And the University of Crete.
For that purpose, the team of the team coding System is based on the chaotic pattern produced when a powerful laser interacts with a small container ethanol as a result of scrambled egg light rays. This process, reinforced by thermal turbulence in the liquid, ensures that the original information is hidden in recognition, making it almost impossible to decode using conventional methods.
For a way to collect the encrypted data, the researchers turned to AI. By training a neural network to recognize and decode the encrypted holograms, they reached an accuracy rate of 90-95% when collecting the original images.
“We came up with the idea of training neural networks to recognize the incredibly fine details of the encrypted light patterns,” Tzortzakis continued.
“By creating billions of complex connections or synapses, we were able to reconstruct the original light beam shapes within the neural networks. This meant that we had a way to make the decoding key that was specific to every configuration of the coding system.”
“The method we have developed is very reliable, even in hard and unpredictable circumstances, with regard to realistic challenges such as hard weather that often limit the performance of optical systems with free space.”
He finally said that “[the team’s] New system reaches an exceptional level of coding by using a neural network to generate the decoding key that can only be made by the owner of the coding system. “
The team has tested the system by coding and decoding thousands of images, including animals, tools, everyday objects and handwritten figures. However. The technology is not entirely ready for commercialization, because the current laser system used in the coding process is extensive and expensive, which limits its usability and affordability.
