Chinese scientists have assured that they have developed an artificial intelligence system (AI) capable of distinguishing between real and false nuclear eyelets, sometimes used as lures. This is the first time in the world that AI is used for this type of arms control, as published by the media .
The research on this system was detailed in a study reviewed by pairs published in April by the Chinese Atomic Energy Institute (CIAE). “China’s position in world negotiations on nuclear disarmament is currently stagnant, but this discovery could help the government improve its position in this regard,” he highlighted the same medium.
The project, a joint initiative proposed by Chinese and American scientists more than a decade ago, has faced three challenges. The first was to try and train the AI system, for which scientists needed access to highly sensitive nuclear data.
Second, the team had to convince Chinese military officers that the system would not run the risk of filtering high secret information. And, the third, they needed to convince other countries – in particular to the United States – to abandon the obsolete verification methods of the Cold War.
Until now, only the first step has taken. “Due to the classified nature of nuclear eyelets and designs of their components, specific data cannot be revealed here,” the CIAE team wrote in its article on atomic energy science and technology.
The AI system, also known as Technical Verification Scheme for deep learning algorithms based on the Interactive Protocol of Zero Knowledgeuse a combination of advanced encryption and nuclear science for its operation.
The researchers executed millions of computer component component simulations, some with real uranium suitable for weapons and other facts to look real using harmless materials such as lead or uranium that are not enriched. In addition, a deep learning system was used to study the movement of neutron through these materials. Over time, AI achieved admirable precision by distinguishing between real and false eyes.
The filtration of ultrasecreta information by AI has been one of the main concerns of this project. To address this problem, the scientists placed a polyethylene wall with 400 small holes between the eyelet and the inspection system, which altered the signals enough to hide the shape of the eyes, while letting the radiation that the AI could use to verify what was looking at.
Stay up to date with the HuffPost! Follow all the news from your mobile in our. You can download it for both.
