Memristor-Based Neuromorphic Computing and Deep Learning

Our primary goal is to advance interdisciplinary research bridging memristive devices and neuromorphic computing. As modern artificial intelligence demands unprecedented computational efficiency, our research focus is shifting from traditional software-centric models toward innovative hardware-intrinsic, brain-inspired systems. Specifically, we develop high-performance memristor-based synapses to enable efficient in-memory and neuromorphic acceleration of various deep learning algorithms. Through this comprehensive cross-layer synergy, we aim to realize highly energy-efficient, next-generation semiconductor hardware and intelligent systems uniquely tailored for the upcoming era of diverse and pervasive artificial intelligence applications.

Next-Generation Semiconductor Devices

The secondary goal is to develop highly efficient and stable semiconductor devices based on next-generation semiconductors such as organic compounds, colloidal quantum dots, and perovskites. These materials offer unique advantages, including high efficiency and broad applicability, yet they significantly lack long-term stability, a critical factor for commercialization. The research focuses on enhancing the long-term stability of these advanced semiconductor devices, including interfacial engineering, device structural engineering, encapsulation engineering, and post-treatment engineering. Through these comprehensive approaches, we aim to develop highly efficient, robust, and commercially viable next-generation semiconductor devices.