English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Resistive memory-based reconfigurable systems constructed by CMOS-RRAMintegration hold great promise for low energy and high throughput neuromorphiccomputing. However, most RRAM technologies relying on filamentary switchingsuffer from variations and noise leading to computational accuracy loss,increased energy consumption, and overhead by expensive program and verifyschemes. Low ON-state resistance of filamentary RRAM devices further increasesthe energy consumption due to high-current read and write operations, andlimits the array size and parallel multiply & accumulate operations.High-forming voltages needed for filamentary RRAM are not compatible withadvanced CMOS technology nodes. To address all these challenges, we developed aforming-free and bulk switching RRAM technology based on a trilayer metal-oxidestack. We systematically engineered a trilayer metal-oxide RRAM stack andinvestigated the switching characteristics of RRAM devices with varyingthicknesses and oxygen vacancy distributions across the trilayer to achievereliable bulk switching without any filament formation. We demonstrated bulkswitching operation at megaohm regime with high current nonlinearity andprogrammed up to 100 levels without compliance current. We developed aneuromorphic compute-in-memory platform based on trilayer bulk RRAM crossbarsby combining energy-efficient switched-capacitor voltage sensing circuits withdifferential encoding of weights to experimentally demonstrate high-accuracymatrix-vector multiplication. We showcased the computational capability of bulkRRAM crossbars by implementing a spiking neural network model for an autonomousnavigation/racing task. Our work addresses challenges posed by existing RRAMtechnologies and paves the way for neuromorphic computing at the edge understrict size, weight, and power constraints.

Multi-level, Forming Free, Bulk Switching Trilayer RRAM for Neuromorphic Computing at the Edge