The Gehring lab received 300kEUR via an INCENTIVE GRANT FOR SCIENTIFIC RESEARCH (MIS) from FNRS for the project "TopoBrain".
The human brain outperforms state-of-the art computers when performing tasks involving real-time processing of unstructured sensory data, such as image, video or voice recognition. The reason for this supremacy is that computing in the brain is massively parallel compared to the mostly linear process flow in computers. Emulating such complex networks with conventional CPU is possible but requires an excessive amount of circuits, which consumes so much energy that running future artificial intelligence by conventional CPU is not feasible. Thus, the only solution to this problem are energy-efficient hardware components that emulate neurons and synapses.
The aim of this project is to gain deep understanding of the physics of topological materials (TMs) and to use this understanding to build a memristor device which can mimic the function of a biological synapse.
The results will teach us about the fundamental properties of topological matter. It will inspire new ways to reduce the power needed to manipulate the resistive state of a memristor and thus to perform logic operations. This is the precondition for realizing an artificial brain with synapse numbers approaching that of biological brains.
Comments