What is it about?
We introduce a novel resistance control method, a feedforward pulse scheme that enhances resistance configuration precision and increases the number of programmable levels of memristor devices. High resolution of 512 resistive states with excellent linearity was achieved by employing an optimized write pulse scheme coupled with a two-dimensional topology for precise dopant ion control in TiO2-x (titanium oxide) memristor.
Featured Image
Photo by Igor Omilaev on Unsplash
Why is it important?
The method enables high resolution analog control of memristors with greater accuracy and efficiency irrespective of physical limitation of ON/OFF resistance ratio. Our approach achieves 95.5% accuracy on ResNet-34 with over 20 million parameters through weight transfer, thereby demonstrating the potential of analog memristors in AI model inference. The findings pave the way for future advancements in increasing resistance states, which will enable more complex AI tasks and enhance the in-memory computational capabilities required for AI edge applications.
Read the Original
This page is a summary of: Enhancing memristor multilevel resistance state with linearity potentiation via the feedforward pulse scheme, Nanoscale Horizons, January 2025, Royal Society of Chemistry,
DOI: 10.1039/d4nh00623b.
You can read the full text:
Contributors
The following have contributed to this page
