WiMi Hologram Cloud Inc. (NASDAQ: WiMi) has announced ongoing research into the utilization of machine learning models to optimize operational parameters within Twin-Field Quantum Key Distribution (TF-QKD) architectures. The technical initiative aims to leverage the non-linear fitting and generalization capabilities of neural networks to predict optimal system configurations. By substituting traditional multi-variable Local Search Algorithms (LSA) with pre-trained regression models, the computational overhead required to calculate dynamic hardware parameters is reduced by multiple orders of magnitude. This minimization of latency is designed to accelerate active secret key generation rates and improve real-time adaptability over fluctuating fiber-optic channels.
[ WiMi Neural Network Evaluation Matrix ]
BPNN (Backpropagation) ──► Simplest topology, fastest calculation velocity; built for rapid-response networks.
RBFNN (Radial Basis) ──► Uses hidden-layer kernel functions; optimized for high-dimensional precision.
GRNN (Gen. Regression) ──► Probability density estimation; handles sparse sample data and channel noise.
The company’s research division trained and benchmarked three distinct neural network topologies to assess their predictive precision within complex parameter spaces. The Backpropagation Neural Network (BPNN) demonstrated the highest execution velocity due to its structural simplicity, positioning it as an ideal match for rapid-response systems with moderate precision boundaries. Conversely, the Radial Basis Function Neural Network (RBFNN) and the Generalized Regression Neural Network (GRNN) achieved superior predictive accuracy when managing high-dimensional variables and signal uncertainties. These specialized models allow TF-QKD terminal stations to automatically counter environmental polarization shifts and phase drift by dynamically adjusting intensity and phase parameters at acceptable computational costs.
Moving forward, the Beijing-headquartered holographic and semiconductor technology provider plans to scale its machine learning research into deep learning and reinforcement learning paradigms to accommodate increasingly complex multi-party quantum communication protocols. By integrating these trained software algorithms directly with commercial quantum hardware platforms, WiMi intends to transition its optimization models from simulated test environments into functional, high-rate quantum secured data transport networks.
The official technology announcement, algorithmic parameters, and corporate performance statements can be reviewed here.
July 1, 2026
