Hybrid memristor/ ferroelectric capacitor based memory enables energy-efficient AI that can autonomously learn and make accurate predictions

https://techxplore.com/news/2025-10-memristor-ferroelectric-memory-energy-efficient.html

"share similar structural stack... memristors apply trained models, offering energy efficiency during read operations/ supporting in-memory computing... ferroelectric capacitors: store information via electric field switchable reversible ferroelectric polarization, endurance/ extremely low energy consumption during programming... integrates silicon-doped HfO₂ (of FeCAPs) with titanium scavenging layer (of memristors)... efficient machine learning, stores analog weights, edge AI without forgetting prior knowledge"

Related:

Breaking the memory–computing divide: Ferroelectric memristors unlock new possibilities
https://www.eurekalert.org/news-releases/1134911

Novel method lets multimodal AI update knowledge without losing earlier information

https://techxplore.com/news/2026-07-method-multimodal-ai-knowledge-earlier.html