a Scalable Quantum Neural Network (SQNN) Technology Based on Multi-Quantum-Device Collaborative Computing

From News Desk

WiMi, a Hologram Augmented Reality (“AR”) Technology provider, has announced the development of a Scalable Quantum Neural Network (SQNN) technology based on multi-quantum-device collaborative computing. This technology uses multiple small quantum devices as quantum feature extractors, which extract local features from input data in parallel. The extracted local features are then aggregated into a quantum predictor through classical communication channels to accomplish the final classification task.

This technology aims to overcome the limitations of current quantum computing hardware by enabling multiple small quantum devices to work collaboratively, thereby building an efficient and scalable quantum neural network system. The technology not only achieves classification accuracy comparable to traditional Quantum Neural Networks (QNN) in theory but also introduces a novel approach to optimizing the utilization of quantum computing resources and enhancing data efficiency.

The core architecture of WiMi’s SQNN system consists of three main components –

Quantum Feature Extractor – The quantum feature extractor is responsible for extracting local features from input data. Each quantum device can independently perform feature extraction tasks using Variational Quantum Circuits (VQC) to encode and transform input data. Since these devices operate independently, they can flexibly adapt to quantum devices of different sizes. For instance, larger quantum devices can handle more complex data patterns, while smaller quantum devices can process simpler local features.

Classical Communication Channel – In the SQNN framework, quantum feature extractors transmit the extracted local features to a central computing node via a classical communication channel. This communication process is similar to the concept of Federated Learning, where different computing units process data independently, but the final decision-making process relies on the integration of global information.

Quantum Predictor- The quantum predictor serves as the core computational unit of the entire SQNN system. It receives feature information from multiple quantum feature extractors and performs the final classification decision using quantum circuits. The quantum predictor can employ more complex quantum circuits to optimize classification accuracy and dynamically adjust its computational approach based on the scale of the data.

The technical implementation of WiMi’s SQNN involves the following steps –

Data Preprocessing and Quantum Encoding – Before entering the quantum system, input data undergoes classical preprocessing operations such as standardisation and dimensionality reduction. The data is then mapped to quantum states using encoding methods such as Amplitude Encoding or Angle Encoding.

Sub-feature Extraction – Each quantum device performs independent feature extraction tasks using Parameterized Quantum Circuits (PQC) to transform features and generate local feature representations.

Feature Aggregation and Classification – The output of quantum feature extractors is transmitted to a central node via a classical communication channel. The quantum predictor then aggregates the features and performs the final classification task.

Parameter Optimisation and Training – SQNN employs Variational Quantum Optimisation for training. A classical optimizer, such as gradient descent, is used to adjust the quantum circuit parameters to minimise classification error.

WiMi’s Scalable Quantum Neural Network (SQNN) is said to provide an innovative solution for quantum machine learning, enabling multiple small quantum devices to work collaboratively for efficient classification tasks. Experimental results indicate that SQNN offers strong computational performance and scalability, laying a solid foundation for the integration of quantum computing and artificial intelligence. As quantum hardware continues to advance, SQNN is expected to become a crucial component of large-scale quantum machine learning systems, driving revolutionary changes in AI and data science.

Disclaimer – The details expressed in this post are from the companies responsible for sending this post for publication. This website doesn’t endorse the details published here. Readers are urged to use their own discretion while making a decision about purchasing or using a product/service related to this company, or using this information in any way. There has been no monetary benefit to the Publisher/Editor/Website Owner for publishing this post and the Website Owner takes no responsibility for the impacts of purchasing or using these products/services on the reader, or using this information in any way.