Chips that can simulate the brain have been tested in creating a neural network with memory

The artificial neural network developed by Graz scientists consumes 4-16 times less energy than analogues in classical computing, as it works on neuromorphic chips, the mechanism of which is similar to the work of biological neurons. The network can remember the data needed to find connection between objects, which is also reflected in its capacity for self-learning. What's more, the calculations are performed in-memory, helping not only to reduce power consumption, but also to reduce latency. The results of the study can be found in the journal Nature Machine Intelligence.