Yigit Demirag

Biography

Yigit obtained his BS and MS degrees in Electrical and Electronics Engineering from Bilkent University, Ankara. He is currently pursuing his Ph.D. degree at the Institute of Neuroinformatics (INI) of ETH Zurich and the University of Zurich. During his MS studies, he worked on the multiphysics modeling of phase change memory-based synaptic devices for neuromorphic computing. He was also involved in several research internship positions at Samsung Electronics, EPFL, and CERN. He is currently focusing on developing neuroscience-inspired learning algorithms and their implementations on hybrid CMOS-memristor architectures. His primary motivation is to accelerate the development of AGI by realizing perfectly optimized hardware for very large neural networks.

Recurrent networks of spiking neurons with memristive synapses

The aim of this project is to realize a hybrid memristive-CMOS neuromorphic event-based neural processing system that can extract features and classify temporal data streams. In particular, the project aims to start by employing a recently developed neuromorphic chip with integrated memristive devices to characterize the integrated memristor properties, perform preliminary spiking neural network experiments applied to spatio-temporal patterns, and implement spiking recurrent models to achieve robust spatio-temporal pattern classification. The main part of the research is performed at University of Zurich, Switzerland. Additionally, Yigit is spending the secondments at Forschungszentrum Jülich, Germany for the characterization of stochastic switching in redox-based memristive devices and derivation of specifications for CMOS interfacing circuits (3 Months) and at IBMResearch – Zurich, Switzerland for the development of specifications, and design of interfacing CMOS circuits for the construction of neuromorphic computing CMOS architectures with memristive devices (3Months).

The University of Zurich (UZH) is a Cantonal university dedicated to higher learning and research. UZH is one of the leading research universities in Europe, devoted to both scientific research and teaching. Two thousand lecturers in 140 special institutes provide the broadest range of subjects and courses available from any Swiss seat of higher education. With 24,000 students and 1,900 graduates every year, Zurich is also Switzerland’s largest university.

The Institute of Neuroinformatics (INI) is a joint institute of the Swiss Federal Institute of Technology Zurich (ETHZ) and the University of Zurich (UZH). It was founded in 1995 to carry out experimental, theoretical and applied research with the aim of discovering key principles by which the brain is built and works, and to use this knowledge in practical applications where possible.

Publications:

Analog weight updates with compliance current modulation of binary ReRAMs for on-chip learning

PCM-trace: Scalable Synaptic Eligibility Traces with Resistivity Drift of Phase-Change Materials

Reconfigurable halide perovskite nanocrystal memristors for neuromorphic computing

Biologically-inspired training of spiking recurrent neural networks with neuromorphic hardware

Mosaic: in-memory computing and routing for small-world spike-based neuromorphic systems