WiMi Hologram Cloud Inc. (NASDAQ: WiMi) is a global Hologram Augmented Reality technology provider with a new deep convolutional neural network designed for image recognition. The company has developed a model utilizing quantum parameterized circuits as its core computing structure, a departure from traditional deep learning approaches and a step toward harnessing quantum computing’s parallel processing capabilities. This technology addresses limitations in computational complexity, memory consumption, and training efficiency by combining quantum circuits with a quantum-classical hybrid training mechanism. The model uses quantum parameterized circuits as its core computing structure, performs feature extraction on image data through quantum convolutional layers, and utilizes a quantum classification layer to complete the final recognition task. The quantum feature fusion module can reduce some computational overhead compared to traditional methods.
Quantum Parameterized Circuits for Image Recognition
WiMi Hologram Cloud Inc. has demonstrated a quantum deep convolutional neural network achieving progress, signaling a deepening of the application of quantum machine learning in artificial intelligence tasks such as image recognition, and also provides a new research direction for the realization of future large-scale quantum intelligent computing systems. WiMi’s approach centers on a fundamentally different computational structure, quantum parameterized circuits, as the core of its deep convolutional neural network. The new model utilizes these circuits to perform feature extraction on image data through quantum convolutional layers, culminating in a quantum classification layer for final recognition. This architecture mirrors the hierarchical structure of classical deep convolutional neural networks, but crucially, it leverages the parallel computing capability of quantum circuits to enhance processing speed when handling high-dimensional data.
The system begins by mapping classical image data into the quantum state space via a data encoding module, a necessary step as quantum computers operate on quantum state information. This conversion of pixel information into qubit probability amplitudes is achieved through methods like amplitude encoding and angle encoding, preparing the data for quantum circuit processing. Following encoding, the quantum convolutional layer extracts features using parameterized quantum gates, functioning similarly to convolution kernels in classical networks but with the potential for higher computational efficiency. At the heart of this layer are WiMi’s specifically designed circuits, composed of rotation gates, control gates, and entanglement gates.
These gates manipulate the evolution of quantum states through trainable parameters; rotation gates adjust qubit state angles, control gates establish correlations, and entanglement gates create complex quantum entanglement structures. The company explains that “Through these operations, the quantum circuit can form feature extraction capabilities similar to those of classical convolutional layers while possessing higher expressive power.” As these layers stack, the network progressively extracts higher-level image features, starting with edges and textures and culminating in complex shapes and structural information. The parallel nature of quantum states within the computation process allows for exponentially larger state spaces, dramatically improving efficiency. To overcome the limitations of current quantum hardware, WiMi has implemented a quantum-classical hybrid training mechanism. This approach utilizes the quantum circuit for forward computation while relying on classical computers for parameter updates.
This draws on the principles of variational quantum algorithms, combining parameterized quantum circuits with classical optimizers to solve complex problems with limited quantum resources. The quantum feature fusion module integrates feature information from different qubits through additional quantum gate operations, completing information integration through the quantum state evolution process, which can reduce some computational overhead.
Quantum Convolutional Layer Architecture & Feature Extraction
The pursuit of quantum machine learning continues to gain momentum, extending beyond theoretical exploration and into the domain of applied technologies. WiMi Hologram Cloud Inc. is a global Hologram Augmented Reality technology provider. This development is noteworthy as it originates from a company focused on Hologram Augmented Reality, signaling a broadening interest in quantum computing beyond traditional technology and artificial intelligence firms. While many organizations explore quantum algorithms, WiMi’s approach focuses on a concrete architectural implementation, utilizing quantum parameterized circuits as the foundational computing structure for its deep learning model. Central to this architecture is a layered system mirroring classical convolutional neural networks, but adapted for quantum processing. This conversion, achieved through techniques like amplitude or angle encoding, allows the image data to be effectively processed by quantum circuits.
The key advantage lies in the potential for highly parallel feature extraction, leveraging the ability of quantum gates to act on multiple superposition states simultaneously when analyzing complex image structures. This parallel processing within an exponentially large state space promises significant improvements in computational efficiency when handling high-dimensional data. This approach leverages the strengths of both computing paradigms; the quantum circuit handles forward computation, while classical computers manage parameter updates. During training, image data is encoded into quantum states, processed through the circuit, and then analyzed to calculate the error between the network’s output and the true label. Classical optimization algorithms then calculate gradient information and update the trainable parameters within the quantum circuit, forming a collaborative training process. The company states that the quantum feature fusion module can reduce some computational overhead compared to traditional methods, emphasizing the practical considerations driving their design choices. The theoretical potential for computational advantage is substantial, with the model potentially offering exponential acceleration in certain tasks compared to traditional deep convolutional neural networks.
WiMi’s technology demonstrates the potential value of quantum computing in the field of artificial intelligence. By combining the physical properties of quantum computing with the model structures of deep learning, new intelligent systems with higher expressive power and stronger computational efficiency can be constructed.
Data Encoding & Quantum Feature Fusion Processes
WiMi Hologram Cloud Inc. is integrating quantum computing into its deep convolutional neural networks, a move that distinguishes it from many other firms focused solely on traditional machine learning approaches. The company, a global Hologram Augmented Reality technology provider, has achieved progress in developing a quantum system for image recognition, and a crucial element of this advancement lies in how classical image data is translated into a quantum format. Following data encoding, WiMi’s quantum convolutional layer undertakes feature extraction, mirroring the function of kernels in classical convolutional neural networks. However, instead of traditional mathematical operations, this layer utilizes a set of parameterized quantum gates acting on local qubits.
WiMi’s research indicates the ability to perform highly parallel feature extraction due to the superposition of quantum states, significantly improving efficiency when handling complex image structures. The company explains that “These quantum gates can form functions similar to convolution filters, performing feature mapping on the input quantum states.” A key innovation within the network architecture is the quantum feature fusion module, designed to integrate information from different qubits. This module employs quantum entanglement mechanisms to fuse image features from various regions, creating higher-dimensional representations with enhanced discriminative power. Unlike traditional neural networks that rely on matrix multiplication for feature fusion, WiMi’s approach completes information integration through quantum state evolution, which can reduce some computational overhead. The subsequent quantum classification layer then outputs classification results by measuring the probability distribution of the quantum states, a process similar to fully connected layers in classical networks but performed within the quantum state space.
Hybrid Quantum-Classical Training Mechanism for WiMi’s Model
WiMi Hologram Cloud Inc. Central to this advancement is a hybrid training mechanism, a pragmatic approach designed to bridge the gap between current quantum hardware limitations and the demands of complex machine learning tasks. This conversion allows the quantum circuits to process the image data effectively. The heart of the quantum convolutional layer lies in specifically designed circuits composed of basic quantum logic gates, rotation, control, and entanglement, which manipulate quantum states through trainable parameters. Realizing this potential requires addressing the limitations of current quantum hardware.
