Techifab's memristor-based chip processes and stores data directly at the point of origin


"energy efficient, increased computational power, fast, decentralized data analysis... memristors process not just 0 and 1,' but values in between, enabling algorithms simulating neural networks...artificial synapses of bismuth/ iron oxide handle complex computations such as matrix multiplication... current prototype has 32 memristors, next phase more than 200 to model complex neural networks... autonomous vehicles. predictive machine maintenance, industrial sensor systems, edge computing"

Related:

Ultra-small neuromorphic chip learns and corrects errors autonomously

Skia improves efficiency and performance in computer processors
https://www.eurekalert.org/news-releases/1083349

Brain-like chips are boosting computers and battling cybercrime

Low-power memristor for neuromorphic computing: From materials to applications

Comments

Post a Comment