Neural holography enables multi-beam steering for multi-task AI classification
Researchers from CICT and partners say a new holographic neural network can handle multiple image-recognition tasks at once using a single phase-only spatial light modulator. The approach could reduce optical hardware complexity and improve the practicality of compact photonic AI systems.
Why it matters: - The new holographic neural network is designed to let optical AI systems do multiple jobs on one compact platform. - The approach could cut hardware reconfiguration, lower system complexity and make optoelectronic computing more practical for multi-task inference. - The method targets a core bottleneck in photonic AI: how to scale from single-task classification to flexible multi-task processing without adding bulky hardware.
What happened: - Researchers from China Information Communication Technologies Group Corporation, Peng Cheng Laboratory, Wuhan University and Nanyang Technological University published a paper in Opto-Electronic Technology. - The paper is titled “Neural Holography enabled Multi-Beam Steering for Multi-Task Classifier.” - The publication appears in volume 2, issue 3 of the 2026 journal issue, under DOI 10.29026/oet.2026.260019. - The team built an end-to-end Holographic Neural Network that uses a single phase-only spatial light modulator. - The system performs input-phase encoding electronically and diffraction-based computation optically. - A camera outputs the final classification results directly.
The details: - The system uses light propagation to carry out parallel computation in the optical domain. - The design relies on a simple optical architecture rather than wavelength switching, polarization changes or mechanical hardware reconfiguration. - The researchers say the setup can process multiple image-recognition tasks through intelligent beam splitting and steering. - Offline simulation training closely matched experimental inference results because of the low-complexity optical system. - The team used an end-to-end optoelectronic joint training framework. - The optical input layer also exploits the nonlinear characteristics of electrical-to-optical conversion. - Inspired by cross-stitch multi-task networks, the researchers proposed a phase-multiplexing holographic scheme and a joint training algorithm. - A task correlation factor was added during error backpropagation to reduce crosstalk between tasks. - For related tasks such as dual-digit recognition, multi-task training exceeded single-task accuracy. - For different task pairs such as handwritten digit recognition and fashion item recognition, tuning the task correlation factor reduced interference. - The researchers demonstrated synchronous recognition of four tasks with about 90% classification accuracy. - Keywords listed with the paper include photonic neural network, pattern classification, optical wireless communications, intelligent beam steering and optical holography.
Between the lines: - The work is aimed at making photonic AI more scalable by replacing physical multiplexing with algorithmic multiplexing. - That shift matters because optical systems are often limited more by alignment and hardware overhead than by raw computational speed. - The results suggest that task-aware training can improve multi-task performance even in a tightly constrained optical setup. - The paper positions neural holography as a way to unify computation and beam steering inside one device architecture.
What's next: - The research points toward larger multi-task optical classifiers that could handle more tasks without added hardware complexity. - Future work is likely to focus on improving accuracy, expanding task counts and testing the architecture in broader real-world inference settings. - The underlying platform could also support next-generation intelligent beam control and optoelectronic AI processing.
Disclaimer: This article was produced by AGP Wire with the assistance of artificial intelligence based on original source content and has been refined to improve clarity, structure, and readability. This content is provided on an “as is” basis. While care has been taken in its preparation, it may contain inaccuracies or omissions, and readers should consult the original source and independently verify key information where appropriate. This content is for informational purposes only and does not constitute legal, financial, investment, or other professional advice.
Sign up for:
China Arts & Entertainment Daily
The daily local news briefing you can trust. Every day. Subscribe now.
Check Your Email!
We sent a one-time activation link to: .
Confirm it's you by clicking the email link.
If the email is not in your inbox, check spam or try again.
Welcome back!
is already signed up. Check your inbox for updates.