HomeScienceMemristor Arrays Enable Optical Reserv
SCIENCE

Memristor Arrays Enable Optical Reservoir Computing

Researchers developed an optically controlled reservoir computing system using oxide memristors that improves processing accuracy and cuts hardware needs.

WHAT YOU NEED TO KNOW
  • Oxide memristor arrays with wavelength-dependent bipolar photoresponses enabled an optically controlled reservoir computing system.
  • Bipolar coding demonstrated higher accuracy than unipolar coding in word recognition and time-series prediction tasks.
  • Parallel coding allowed multi-source signal fusion in a single physical reservoir, cutting hardware consumption.
  • The paper was published in Nature Communications on July 31, 2026.

Researchers built an all-optically controlled physical reservoir computing system using oxide memristor arrays, according to a paper published in Nature Communications. The system uses wavelength-dependent bipolar photoresponses to overcome accuracy limitations found in traditional optoelectronic reservoir computing setups.

Conventional optoelectronic reservoir computing platforms combine electronic and photonic computation, but their devices typically rely on unipolar photoresponses. That restriction limits reservoir state diversity and computational accuracy. To address this, researchers from institutions including the Ningbo Institute of Materials Technology and Engineering developed uniform oxide memristor arrays that exhibit a bipolar photoresponse depending on light wavelength. The response stems from light-induced dynamic evolution of oxygen vacancies within the devices.

Light control and coding

By tuning the power density and irradiation mode of dual-wavelength light, the research team achieved dynamic control over photocurrent relaxation and nonlinearity. They used these properties to create bipolar and parallel coding strategies. In tests involving word recognition and time-series prediction tasks, the bipolar coding strategy demonstrated higher accuracy than unipolar coding.

The parallel coding strategy supported multi-source signal fusion within a single physical reservoir. Nature Communications reported that this approach maintained high computational accuracy while significantly reducing hardware consumption. Lingxiang Hu and Dian Jiao contributed equally to the research work.

The project received funding from the National Natural Science Foundation of China, the Zhejiang Provincial Natural Science Foundation of China, the Ningbo Municipal Science and Technology Innovation Yongjiang 2035 Key R&D Plan, and the Ningbo Global Innovation Center of Zhejiang University. The manuscript was submitted on January 29, 2026, and accepted on July 20, 2026.

Xentir Media
Xentir Media NewsroomSource-backed AI and technology coverage, drafted by Xentir's automated editorial system under fixed human-set rules. See our editorial policy and AI usage policy.
J
Jomon · Founder & EditorFounder and editor of Xentir Media. Sets the editorial rules the newsroom system runs under, and is accountable for its corrections. About Jomon · [email protected]
The Xentir Brief
The developments worth knowing — one useful email.
Get the Brief →