Duygu Kuzum

Our research is at the interface of electronics and neuroscience. We develop nanoelectronic devices beyond CMOS for energy efficient neuro-inspired computing. We investigate algorithm-device codesign approaches and emerging nonvolatile device technologies such as phase change memory, resistive change memory, and magnetic random-access memory to build hardware accelerators for neural networks. We harness nonlinear switching characteristics of quantum materials to implement neural network computation in hardware. We develop a broad range of implantable device technologies for the brain. We apply innovations in nanoelectronics to develop new neurotechnologies, which will help to better understand circuit-level computation in the brain. We explore 2D materials to develop optically transparent neural implants which can record brain signals with high spatial and temporal resolution through combination of electrical and optical recording techniques. We work on model-based and machine learning based computational methods to analyze neural recordings from the brain. Our research at the interface of electronics and neuroscience aims to better understand computation and information processing in the brain and translate that knowledge into new learning systems with brain-like capabilities.