Memristors And Their Application In NEUROMORPHIC COMPUTING

Venturing into memristor technology brings us nearer to brain-like Al. This leap could reshape how we think about computing.

- DURAIARASU E

Memristors, a portmanteau of ‘memory’ and ‘resistor,’ are a groundbreaking class of passive two-terminal circuit elements. First theorised by Leon Chua in 1971 and physically realised by HP Labs in 2008, memristors retain a state of resistance based on the history of voltage and current passed through them. This unique property makes them similar to synapses in the human brain, enabling them to play a crucial role in neuromorphic computing.

Fundamental principles of memristors

At their core, memristors operate on the principle that their resistance can change when an electric current is applied. This resistance remains until another current is applied to change it again, thanks to the movement of ions or vacancies within the memristor’s material, typically a thin film of metal oxide.

Key characteristics of memristors include their non-volatility, allowing them to retain their resistance state even when the power is off, which is essential for memory storage. They are also highly scalable, with the ability to be manufactured at nanoscale dimensions, facilitating high-density integration. Moreover, memristors consume less power compared to traditional memory devices due to their efficient operation.

Neuromorphic computing: An overview