Pillar 4 - Memory Centric Research
Memory-Centric Research Publications
In Year 4, Virginia Tech contributed six peer-reviewed publications to the UPWARDS memory-centric research effort, spanning near-memory and in-memory computing, MRAM- and SRAM-based compute-in-memory architectures, hyperdimensional computing, and binary neural network accelerators in advanced CMOS and FDSOI technology nodes. The work emphasizes energy efficiency, scalability, and security at the edge.
Each project is summarized on the dedicated research projects page, including design specifications, key research contributions, and representative figures. A full reference list with publication links is included at the bottom of that page.
UPWARDS Lightning Talks
- Virginia Tech will organize the upcoming UPWARDS Lightning Talks session in June 2026, focusing on the theme “Semiconductor and Circuit Design for Artificial Intelligence.” (Coordinated by Yang (Cindy) Yi).
- This session will highlight cutting-edge advances in circuit- and system-level design enabling next-generation AI hardware, with an emphasis on energy efficiency, scalability, and emerging memory-centric computing paradigms.
- The event aims to bring together researchers from the U.S. and Japan to exchange ideas and foster collaboration in semiconductor technologies critical to AI acceleration.
Research Facility & Matrix
- Mapped VT Research Strengths
- Identified 9 key semiconductor research areas aligned with UPWARDS taxonomy
- Highlighted strengths in CIM, neuromorphic systems, lithography, AI, and heterogeneous integration
- Documented Facilities & Equipment
- Submitted 49 major instruments across cleanroom and advanced labs
- Covered full capabilities in fabrication, characterization, and prototyping (FPGA/neuromorphic)
- Enabled Collaboration & Visibility
- Provided centralized resources to support U.S.–Japan collaboration and student exchange
- Made VT facilities publicly accessible via UPWARDS network website
- Strengthened Network Coordination