Prof. Seongjae Cho of the Department of Intelligent Semiconductor Engineering is conducting research on DRAM-Flash-based neuromorphic computing and Processing-in-Memory (PIM) technologies, which are emerging as key technologies for next-generation AI semiconductors. His research focuses on developing new memory architectures for high-speed, low-power AI semiconductor systems.
The research integrates DRAM and Flash memory, which have different characteristics, within a single AI system. DRAM is utilized as a synaptic device optimized for fast, learning-intensive tasks, while Flash is used for large-scale parameter storage and high-speed inference. The research also explores three-dimensional integration of CMOS, DRAM, and Flash to efficiently combine memory and computing functions within a compact architecture.
This approach aims to overcome the performance limitations of conventional single-memory technologies and enable high-speed, low-power AI semiconductor systems capable of improving both learning and inference performance. In particular, three-dimensional integration of DRAM and Flash, together with PIM technology, is expected to improve AI computing efficiency by reducing data movement between memory and processing units.
Prof. Cho continues to advance research in next-generation memory and AI semiconductor technologies, integrating device, fabrication, circuit, and system-level approaches. His research is expected to contribute to the development of high-performance, energy-efficient semiconductor technologies for the AI era.
Related Link
Read the Ewha News Article (KR): 지능형반도체공학전공 조성재 교수, 2026 차세대지능형반도체기술개발사업(소자) 선정
