New memristors are to perform machine-learning tasks more efficiently on conventional computers. They emulate the working principles of neurons in the human brain. These electronic circuits that can be used in a far wider range of applications are currently being developed by researchers at the University of Zurich (UZH), the Swiss Federal Institute of Technology in Zurich (ETH) and the Swiss Federal Laboratories for Materials Science and Technology (Empa).
According to a UZH press release, these memristors no longer need to be configured for one operation mode in advance and can instead easily switch while in use, between a mode in which the signal grows weaker over time and dies (volatile mode), and one in which the signal remains constant (non-volatile mode).
This is similar to the two operation modes also found in the human brain: either stimuli at the synapses start out strong and then gradually become weaker, or new synaptic connections to other neurons form in the brain while we learn. These connections are longer-lasting. “To our knowledge, this is the first memristor that can be reliably switched between volatile and non-volatile modes on demand,” explains Yiğit Demirağ, a doctoral student in Professor Giacomo Indiveri’s group at the Institute for Neuroinformatics of the University of Zurich and ETH Zurich. These memristors come closer to real neurons than previous ones, says Indiveri. Before these memristors can be used in computer technology, they will need to undergo further optimization.