WiMi Studies Hybrid Quantum Neural Network Architecture with Classical-Quantum Synergistic Innovation to Enhance Efficiency and Accuracy in Image Classification

WiMi Hologram Cloud (WIMI) is researching a hybrid quantum neural network (HQNN) architecture that combines classical and quantum techniques to improve efficiency and accuracy in image classification.

Rhea-AI Summary

WiMi Hologram Cloud (WIMI) is researching a hybrid quantum neural network (HQNN) architecture that combines classical and quantum techniques to improve efficiency and accuracy in image classification.

The architecture targets current NISQ-era quantum hardware constraints by fusing classical convolutional layers with parallel parameterized quantum circuits. A classical convolutional block handles feature extraction and dimension reduction, while a hybrid dense part with parallel PQCs and classical fully connected layers performs prediction. Quantum-classical hybrid optimization, including the parameter-shift rule, is used to jointly train quantum gate parameters and classical weights, aiming to speed high-dimensional feature processing and enhance model stability and generalization.

Historical Context

1 past event · Latest: Aug 13

1 event

1. Aug 13

   AI architecture research

   24h Move

   +2.3%

   Previously disclosed H-QNN application to binary MNIST with reported accuracy and training-efficiency claims.

24h Move is the share-price change in the day after each event; other market factors may also have contributed.

Key Terms

quantum machine learning, hybrid quantum neural network, qubit, parameterized quantum circuits, +1 more

5 terms

quantum machine learning technical

"quantum machine learning (QML), as a cutting-edge field"

Quantum machine learning is the use of quantum computers or simulators to run algorithms that find patterns, make predictions, or learn from data, instead of using only ordinary computers. Think of it as trying a new, potentially faster type of engine for data tasks that may solve some problems more efficiently or in a different way. For investors, advances could lower costs, unlock new services, or create advantages for firms that successfully commercialize or apply the technology, affecting valuations and competitive positioning.

hybrid quantum neural network technical

"the hybrid quantum neural network (HQNN)"

A hybrid quantum neural network is a computing model that combines quantum circuits with classical neural network layers so some calculations run on quantum hardware (or simulators) while others run on ordinary processors. Think of it like a hybrid car that uses an electric motor for certain tasks and a gasoline engine for others; the mix aims to exploit quantum devices for parts of a problem where they may be faster or offer different behaviors. Investors track the idea because it represents a potential route for firms to develop advanced machine‑learning or optimization capabilities that could change computing performance and product features.

qubit technical

"each qubit to carry multiple state information simultaneously"

A qubit is the basic unit of information used in quantum computers, like a coin that can be heads, tails or both at once until you look; this lets quantum machines process many possibilities simultaneously. For investors, qubits matter because their number, quality and stability determine how powerful a quantum computer can be, affecting which companies might gain an edge in fields such as cryptography, drug discovery, materials design or complex financial modeling.

parameterized quantum circuits technical

"through the introduction of parameterized quantum circuits (PQCs)"

Parameterized quantum circuits are programmable sequences of quantum operations with adjustable knobs (parameters) that you tune to make the circuit produce a desired output, similar to adjusting equalizer sliders to improve sound. Investors care because these circuits are the leading approach for running useful tasks—like pattern recognition, optimization and simulation—on current and near-term quantum hardware, so progress can signal commercial potential, partnerships, and value drivers in quantum-focused firms.

parameter-shift rule technical

"such as the parameter-shift rule"

A parameter-shift rule is a practical method used in quantum computing to measure how the output of a quantum circuit changes when a tunable setting is adjusted, by running the circuit with slightly different knob positions and comparing results. For investors, it matters because this technique determines how efficiently quantum algorithms and devices can be trained and optimized; faster, more reliable tuning speeds development and improves the commercial prospects of quantum-based products and services.

AI-generated analysis. How Rhea-AI works. Not financial advice.

See more from StockTitan in Google Search and AI answers. Adds StockTitan as a preferred source · opens Google

Add on Google Not now

BEIJING, Sept, 23, 2026 /PRNewswire/ -- WiMi Hologram Cloud Inc. (NASDAQ: WiMi) ("WiMi" or the "Company"), a leading global Hologram Augmented Reality ("AR") Technology provider, announces that it is researching the application of quantum machine learning to image classification. Its researched hybrid quantum neural network architecture, through synergistic innovation between classical and quantum approaches, will enhance the efficiency and accuracy of image classification. The fusion of quantum computing and machine learning provides technical support for breakthroughs in image classification technology. The superposition property of quantum mechanics enables each qubit to carry multiple state information simultaneously, while the entanglement property realizes non-local correlations between qubits. The combination of these two grants quantum algorithms a natural parallel computing capability, allowing for an exponential reduction in computational complexity when processing high-dimensional data. Quantum machine learning (QML), as a cutting-edge field that integrates quantum mechanical principles with classical machine learning, is expected to significantly reduce computational power consumption while maintaining or even improving classification accuracy by constructing quantized feature mapping and processing mechanisms. Among them, the hybrid quantum neural network (HQNN), by virtue of its synergistic optimization of classical architectures and quantum algorithms, has become the most promising and practically deployable technical path in the NISQ era—it avoids the dependence of fully quantum neural networks on large-scale fault-tolerant quantum hardware, while fully leveraging the parallel advantages of quantum computing through the introduction of parameterized quantum circuits (PQCs).

Current NISQ-era quantum hardware faces technical constraints such as limited number of qubits and prominent noise interference, making it difficult for purely quantum circuits to process large-scale real-world image data. The hybrid quantum neural network architecture researched by WiMi, through the innovative fusion of classical convolutional layers and parallel quantum layers, constructs a hybrid computing system adapted to NISQ devices, achieving precise complementarity between the advantages of classical computing power and quantum parallel capabilities. The core design logic of this architecture lies in modular decomposition, which not only addresses the processing capability limitations of quantum hardware but also maximizes the efficiency advantages of quantum computing. The architecture consists of two major core modules: the classical convolutional block and the hybrid dense part, forming a complete optimization chain from data preprocessing to prediction output. The classical convolutional block undertakes the core tasks of feature extraction and dimension compression. Through multi-layer convolution operations and pooling mechanisms, it performs layer-by-layer feature abstraction on the input image data, filters out redundant information, and retains key semantic features, ultimately transforming high-dimensional image data into low-dimensional feature vectors suitable for processing by the quantum layer. This design not only resolves the input dimension limitations of NISQ devices but also, by leveraging the mature feature extraction capabilities of classical convolution, lays a high-quality data foundation for subsequent quantum layer processing, avoiding the computational waste and noise amplification problems that would arise from quantum circuits directly processing raw image data.

Adopting the synergistic design of classical fully connected layers and parallel parameterized quantum circuits (PQCs), it undertakes the model's final prediction task. Unlike traditional serial quantum layer designs, WiMi enables multiple quantum circuits to synchronously execute feature processing tasks through parallelized PQC deployment, leveraging the quantum superposition and entanglement properties to achieve parallel computation of multiple groups of features, significantly reducing the overall computation time. The implementation of synergistic training between the quantum layer and the classical fully connected layer is achieved through quantum-classical hybrid optimization strategies such as the parameter-shift rule, which feeds back the measurement results of the quantum circuits to the classical optimizer, realizing joint iterative optimization of quantum gate parameters and classical layer weights, thereby improving the model's convergence stability and generalization capability.

Compared to traditional classical CNNs and single quantum models, the parallel PQC design and classical-quantum synergistic processing mechanism, while reducing the model parameter scale, significantly enhance the processing speed of high-dimensional features, making it particularly suitable for real-time classification scenarios involving massive image data. In addition, the feature preprocessing capability of classical convolutional layers complements the noise resistance characteristics of quantum circuits, enabling the model to maintain stable classification performance even in complex data environments and under quantum hardware noise interference. Beyond that, the modular design of the architecture allows flexible adaptation to NISQ devices of different scales. With the iteration of quantum hardware technology, performance can be continuously upgraded by increasing the number of qubits and optimizing the PQC structure.

The hybrid quantum neural network architecture researched by WiMi not only provides a practically feasible solution for breaking through the computational power bottleneck in image classification technology, but also demonstrates its technological foresight in the field of quantum machine learning. In the future, WiMi will continue to deeply cultivate innovations in quantum machine learning algorithms and architectures, harnessing the breakthrough power of quantum technology to lead the artificial intelligence industry into a new era of efficient computing.

About WiMi Hologram Cloud

WiMi Hologram Cloud Inc. (NASDAQ: WiMi) focuses on holographic cloud services, primarily concentrating on professional fields such as in-vehicle AR holographic HUD, 3D holographic pulse LiDAR, head-mounted light field holographic devices, holographic semiconductors, holographic cloud software, holographic car navigation, metaverse holographic AR/VR devices, and metaverse holographic cloud software. It covers multiple aspects of holographic AR technologies, including in-vehicle holographic AR technology, 3D holographic pulse LiDAR technology, holographic vision semiconductor technology, holographic software development, holographic AR virtual advertising technology, holographic AR virtual entertainment technology, holographic ARSDK payment, interactive holographic virtual communication, metaverse holographic AR technology, and metaverse virtual cloud services. WiMi is a comprehensive holographic cloud technology solution provider. For more information, please visit http://ir.wimiar.com.

Translation Disclaimer

The original version of this announcement is the officially authorized and only legally binding version. If there are any inconsistencies or differences in meaning between the Chinese translation and the original version, the original version shall prevail. WiMi Hologram Cloud Inc. and related institutions and individuals make no guarantees regarding the translated version and assume no responsibility for any direct or indirect losses caused by translation inaccuracies.

View original content:https://www.prnewswire.com/news-releases/wimi-studies-hybrid-quantum-neural-network-architecture-with-classical-quantum-synergistic-innovation-to-enhance-efficiency-and-accuracy-in-image-classification-302887649.html

SOURCE WiMi Hologram Cloud Inc.

FAQ

AI-generated questions and answers. How Rhea-AI works. Not financial advice.

What are the main components of WiMi's hybrid quantum neural network architecture?

The architecture consists of two core modules: a classical convolutional block and a hybrid dense part