WiMi explores shallow hybrid quantum-classical neural network for image classification

BEIJING - WiMi Hologram Cloud Inc. (NASDAQ:WIMI), a technology company with a market capitalization of $371 million and annual revenue of $61 million, announced Thursday it is exploring a shallow hybrid quantum-classical convolutional neural network (SHQCNN) model for image classification applications.

The company's research focuses on implementing enhanced variational quantum methods that transform quantum state optimization problems into classical optimization problems. According to the press release, these methods incorporate complex quantum gate combinations and parameterized forms to better represent quantum features of image data.

The SHQCNN model utilizes kernel encoding in the input layer to map image data from low-dimensional to high-dimensional feature space, potentially improving data distinction. For the hidden layer, the model employs variational quantum circuits with a more streamlined structure compared to traditional deep neural networks.

In the output layer, WiMi implements a mini-batch gradient descent algorithm that selects small batches of training data for each iteration rather than using the entire dataset, which the company states allows for faster computation and better convergence.

WiMi Hologram Cloud, which describes itself as a holographic cloud technology provider, offers services including in-vehicle AR holographic HUD, 3D holographic pulse LiDAR, holographic semiconductors, and metaverse holographic AR/VR devices.

The announcement comes as quantum computing research continues to expand across various industries, with companies exploring hybrid approaches that combine quantum and classical computing methods to address practical challenges in fields like image processing. According to InvestingPro analysis, WIMI maintains strong financial health with a current ratio of 3.5 and trades at an attractive price-to-book ratio of 0.24, though the stock has declined 58% year-to-date. InvestingPro analysis suggests the stock is currently undervalued, with 8 additional key insights available to subscribers.

In other recent news, WiMi Hologram Cloud Inc. announced several advancements in quantum computing technology. The company developed a single-qubit quantum neural network aimed at improving computational efficiency in artificial intelligence by utilizing high-dimensional quantum bits for multi-class classification tasks. Additionally, WiMi introduced Quantum Semi-Supervised Learning technology to enhance machine learning with limited labeled data, leveraging quantum computing's parallel processing capabilities. The company also began exploring Quantum Dilated Convolutional Neural Networks to address challenges in processing complex data by integrating quantum computing into traditional neural network architecture. Furthermore, WiMi is investigating a quantum image encryption algorithm using four-dimensional chaos, which encodes classical images into quantum information. The company is also researching a quantum crypt generator, QryptGen, which uses quantum machine learning to generate encryption keys with enhanced security. These developments reflect WiMi's ongoing efforts to harness quantum computing for various technological advancements.

This article was generated with the support of AI and reviewed by an editor. For more information see our T&C.