This course aims to study intelligent semiconductor technologies for the development of hardware systems optimized for artificial intelligence (AI) computation. AI technology is advancing rapidly and significantly transforming human life. Through neural network–based machine learning, which mimics how the human brain transmits and processes information via neurons and synapses, performance levels that considerably exceed those of traditional computing systems have been achieved in applications such as natural language processing and image recognition. This advancement is generating substantial ripple effects across various industries and society at large.
Nevertheless, the CPU and GPU architectures founded upon the von Neumann model—the cornerstone of modern computing—are hindered by memory bottlenecks. Consequently, training large-scale neural networks often requires days or even weeks utilizing data center–level infrastructures and expends substantial amounts of energy. To address these constraints, extensive research is ongoing regarding intelligent semiconductor technologies designed to supplant or complement existing hardware optimized for traditional computing. This course investigates the theories, fabrication processes, and applications of neuromorphic synaptic devices, which represent a new frontier developed to surpass the limitations of conventional semiconductor technologies that have evolved over several decades. Students will acquire knowledge of current technological trends in low-power AI computation at the hardware level and will discuss prospective future directions for progress.

For POSTECH Undergraduates,
EECE211 (반도체전자공학I)

--

For Students Outside POSTECH,
Required Undergraduate-Level Semiconductor Courses

--

Please contact the instructor if anything is unclear.

None

반도체 산업

인공지능 기술의 발전과 기존 von Neumann 아키텍처

트랜지스터와 주요 메모리 소자

반도체 회로에서의 인공지능 연산

기존 메모리를 이용한 인공지능 반도체

Resistive 인공지능 반도체

Magnetic 인공지능 반도체

Ferroelectric 인공지능 반도체

Capacitor-based 인공지능 반도체

- 국내 여러 대학에 공유되는 수업이어서 몇가지 필요한 행정 절차 진행을 고려해서 3/3 (화) 첫 강의 업로드 될 예정입니다.

- 비대면 비실시간 온라인 강좌 (VOD) 로 운영됩니다.

- EECE490D & DISU431 은 완전히 동일한 과목입니다. 대학 내부 행정 절차 때문에 2개 과목이 cross listing 되어 있는데, 수강 신청은 DISU431 로 진행해 주세요.

- 매주 월요일 강의가 업로드 되고, 돌아오는 일요일까지 자유롭게 시청하면 됩니다.