Shares of WiMi Hologram Cloud Inc. (WIMI) rose 3.03%, a gain of $0.04, to $1.37 during Thursday's session after the company unveiled a new quantum computing design centered on a hybrid quantum-classical model. WiMi said the technology targets classical data classification tasks directly, and the stock held its gains through steady midday trading following the quantum machine learning reveal.

Quantum Neural Network Breakthrough

WiMi Hologram Cloud introduced a quantum convolutional neural network built around three-qubit interaction layers. The design aims to boost expressive power and entanglement generation in quantum models, and company researchers described the structure as marking a shift toward multi-body interaction models.

The new interaction layer sits alongside the network's standard convolutional units. Working together, these components extract features from input data while limiting circuit depth and complexity — a balance that WiMi said keeps the system hardware-friendly and practical. That trade-off is central to quantum machine learning, a research field that applies quantum circuits to machine-learning tasks: today's quantum processors remain noisy, limited-qubit devices, so designs that generate useful entanglement at shallow circuit depths are the ones feasible to run on existing hardware.

According to the company's researchers, three-body interactions significantly expand the reachable quantum state space. That expansion addresses a common expressivity limitation found in earlier quantum neural networks, and as a result the model can capture more complex data patterns efficiently. Whether gains of this kind translate into an advantage over classical neural networks on practical datasets remains an open research question across the field.

Technical Design

The architecture uses a hybrid quantum-classical approach to process information. Classical data is first mapped into quantum states through an encoding strategy: image data relies on block partitioning, while one-dimensional data uses angle encoding. According to the company, this encoding step allows classical information to be represented and processed within the quantum network.

Entanglement analysis showed that the three-qubit layer generates strong, multi-scale entanglement at shallow circuit depths. WiMi said this capability helps the network reliably capture nonlinear correlations within input data and also supports stability when noise is present.

Training combines classical optimizers with quantum circuit parameters through joint iteration. Engineers adjusted initialization strategies to resolve gradient vanishing and instability issues, and the company reported that the model consequently converges steadily across both binary and multi-class classification tasks.

Company Background and Next Steps

WiMi Hologram Cloud operates as a global provider of hologram augmented reality technology, and the company has recently expanded its research scope to include quantum computing applications, releasing a series of research announcements in the area. The latest release builds on that broader technical strategy and roadmap, and it describes work at the research stage: the company positions deployment on real quantum devices as a future step rather than a current offering.

WiMi said it plans to scale the model for larger applications. Future targets include higher-dimensional image data and complex time-series analysis, while cross-modal data fusion remains a key area for further development.

The company said it will continue testing noise robustness and hardware compatibility going forward. Those efforts aim to move the model from theory toward deployment on real quantum devices, a progression WiMi said supports its long-term quantum computing ambitions.

This article first appeared on Blockonomi.