A memristor is a two-terminal electronic component that acts like a resistor, but unlike a traditional resistor, it can "remember" the amount of charge that has previously flowed through it. This memory capability arises from the component's ability to change its resistance based on the history of current flow. It's considered the fourth fundamental circuit element, alongside the resistor, capacitor, and inductor.

Neural networks based on memristive devices have the ability to improve throughput and energy efficiency for machine learning and artificial intelligence. Memristive-switching devices are known for their relatively large dynamical range of conductance, which can lead to a large number of discrete conductance levels.

Here, 2,048 conductance levels were achieved in a memristor which is more than an order of magnitude higher than previous demonstrations. The root cause of conductance fluctuations in memristors was identified through experimental and theoretical studies and an electrical operation protocol was devised to denoise the memristors for high-precision operations.

These results not only provide crucial insights into the microscopy picture of the memristive switching process but also represent a step forward in commercializing memristor technology as hardware accelerators of machine learning and artificial intelligence for edge applications.