Artificial synapses are not affected by imperfection
A trio of Chinese researchers has developed a platform for artificial intelligence that is far more robust than current architectures. They used a strategy to mimic the encoding information in our brains. The approach has not yet been implemented in the laboratory, but it exploits the non-uniformity that is inherent in artificial neurons as a result from defects in magnetic materials.
Zhe Yuan and Ya Qiao, from the Department of Physics and Center for Advanced Quantum Studies at Beijing Normal University, conducted this research.