Neuromorphic computing using analog memory (AI hardware)

• AI inference using output current from multi-voltage pulse on crossbar array memory architecture 

• Clock-synchronous data processing for fast throughput

• Multiple VMM operations using multiple output current lines for low-energy parallel computing

Neural network modeling for data processing (AI application)

• Practical application for health care, motion detection, traffic control, image/video recognition

• Discrete weight optimization for energy-efficient AI accelerator

• Development of SGD for fine-tuning of synaptic weight