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

The rapid advancements in machine learning across numerous industries haveamplified the demand for extensive matrix-vector multiplication operations,thereby challenging the capacities of traditional von Neumann computingarchitectures. To address this, researchers are currently exploringalternatives such as in-memory computing systems to develop faster and moreenergy-efficient hardware. In particular, there is renewed interest incomputing systems based on optics, which could potentially handle matrix-vectormultiplication in a more energy-efficient way. Despite promising initialresults, developing a highly parallel, programmable, and scalable opticalcomputing system capable of rivaling electronic computing hardware stillremains elusive. In this context, we propose a hyperspectral in-memorycomputing architecture that integrates space multiplexing with frequencymultiplexing of optical frequency combs and uses spatial light modulators as aprogrammable optical memory, thereby boosting the computational throughput andthe energy efficiency. We have experimentally demonstrated multiply-accumulateoperations with higher than 4-bit precision in both matrix-vector andmatrix-matrix multiplications, which suggests the system's potential for a widevariety of deep learning and optimization tasks. This system exhibitsextraordinary modularity, scalability, and programmability, effectivelytranscending the traditional limitations of optics-based computingarchitectures. Our approach demonstrates the potential to scale beyond petaoperations per second, marking a significant step towards achievinghigh-throughput energy-efficient optical computing.

Hyperspectral In-Memory Computing with Optical Frequency Combs and Programmable Optical Memories