WiMi Hologram Cloud is exploring multi-dimensional data pooling technology based on variational quantum algorithms. This is stated in the company’s press release.
The approach combines the Quantum Haar Transform, partial quantum measurement, and a hybrid scheme of parameterized quantum circuits and classical optimizers. According to WiMi, this architecture is designed to reduce the dimensionality of high-dimensional data while preserving local features.
Pooling in machine learning is used to reduce data volume while retaining significant features. WiMi claims that the quantum scheme can perform this operation directly on multi-dimensional data without first converting it to a one-dimensional form.
The company identified potential applications in audio data, images, 3D point clouds, and hyperspectral data. WiMi did not provide a timeline for commercial implementation.
WiMi Hologram Cloud is listed on Nasdaq under the ticker WIMI and specializes in holographic cloud and AR solutions.
Back in May, scientists from ETH Zurich developed a method for creating mathematically perfect randomness. The technology addresses the vulnerability of digital systems and cryptography.