The advent of large AI models has worsened the long-standing von Neumann bottleneck issue. A high-speed, high-density embedded memory may provide a possible solution by enabling the entire AI model to be stored on-chip, thus reducing the reliance on external memory. However, such a memory technology is still elusive to-date. The main objective of this PhD research project is to realize such a technology through deploying novel material platforms. You will join our existing project to investigate…

The advent of large AI models has worsened the long-standing von Neumann bottleneck issue. A high-speed, high-density embedded memory may provide a possible solution by enabling the entire AI model to be stored on-chip, thus reducing the reliance on external memory. However, such a memory technology is still elusive to-date. The main objective of this PhD research project is to realize such a technology through deploying novel material platforms. You will join our existing project to investigate…