Characterization of emerging memories (eNVMs): Analysis of emerging devices using electrical, physical measurements, and TCAD simulations to understand their behavior and overcome limits like variability and endurance.
Analysis of in-memory computing architectures: Study of crossbar architectures at the circuit level, focusing on the impact of non-ideal effects (e.g., parasitic resistances) on computational accuracy and efficiency.
Development of hardware-aware techniques: Implementation of non-linear weight mapping and hardware-aware training to adapt neural networks to analog hardware imperfections.
