Oxide based Circuit Design
Embedded Files
a-InGaZnO based Neuron Circuit
a-InGaZnO based Neuron Circuit
We design and propose an IGZO-based neuron circuit. To achieve accurate co-operation with IGZO-based synaptic transistors, the designed parameters of the neuron circuit are optimized through HSPICE. Additionally, we fabricate and measure the a-IGZO-based neuron circuit. Our neuron circuit can be directly integrated on top of other circuits or onto various substrates due to the low-temperature processability of a-IGZO TFT.
a-IGZO ReLU Neuron Circuit for Artificial Neural Networks
a-IGZO ReLU Neuron Circuit for Artificial Neural Networks
a-IGZO Integrate-and-Fire Neuron Circuit for Spiking Neural Networks
a-IGZO Integrate-and-Fire Neuron Circuit for Spiking Neural Networks
Oxide and CMOS Hybrid Circuit Design
Hybrid and CMOS based Circuit Design
Hybrid and CMOS based Circuit Design
Hybrid(IGZO/Si) Integrate-and-Fire Neuron Circuit for Spiking Neural Networks
Hybrid(IGZO/Si) Integrate-and-Fire Neuron Circuit for Spiking Neural Networks
Our laboratory proposes a hybrid (a-IGZO/CMOS) neuron circuit. By leveraging the extremely low leakage current and low mobility of a-IGZO TFT, the hybrid structure achieves significantly low power consumption compared to conventional CMOS-only neuron circuits. The operation and energy efficiency of the hybrid neuron circuit are validated through simulation.
Project sponsors
삼성전자, KEIT(한국산업기술평가관리원), LG Display(엘지디스플레이), SDC(삼성디스플레이)
Google Sites
Report abuse