Hybrid quantum tensor networks for aeroelastic applications

2025-08-07

AI Summary

This work addresses classification and regression tasks for complex time-series data in aeroelasticity. Method: We propose an end-to-end trainable hybrid quantum tensor network framework. It is the first to employ classical tensor networks for time-series pre-encoding and dimensionality reduction, which subsequently informs the architecture design of a trainable variational quantum circuit inspired by tensor network structure—enabling efficient quantum state encoding and feature extraction. Contribution/Results: By synergistically integrating the high-dimensional compression capability of tensor networks with the nonlinear modeling power of quantum circuits, the framework achieves enhanced model expressivity and training stability while maintaining low qubit overhead. Experiments demonstrate high accuracy on aeroelastic binary classification and excellent regression performance for critical discrete state variables, validating its feasibility and potential for modeling real-world physical systems.

Technology Category

• Machine Learning: Quantum Machine Learning
• Intelligent Robots: State Estimation
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Economics, Online Markets and Human Computation: Economic ramifications for generative AI infrastructure and applications

Abstract

We investigate the application of hybrid quantum tensor networks to aeroelastic problems, harnessing the power of Quantum Machine Learning (QML). By combining tensor networks with variational quantum circuits, we demonstrate the potential of QML to tackle complex time series classification and regression tasks. Our results showcase the ability of hybrid quantum tensor networks to achieve high accuracy in binary classification. Furthermore, we observe promising performance in regressing discrete variables. While hyperparameter selection remains a challenge, requiring careful optimisation to unlock the full potential of these models, this work contributes significantly to the development of QML for solving intricate problems in aeroelasticity. We present an end-to-end trainable hybrid algorithm. We first encode time series into tensor networks to then utilise trainable tensor networks for dimensionality reduction, and convert the resulting tensor to a quantum circuit in the encoding step. Then, a tensor network inspired trainable variational quantum circuit is applied to solve either a classification or a multivariate or univariate regression task in the aeroelasticity domain.

Problem

• Applying hybrid quantum tensor networks to aeroelastic problems using QML
• Solving complex time series classification and regression with quantum circuits
• Developing trainable hybrid algorithms for aeroelasticity tasks

Innovation

• Hybrid quantum tensor networks for aeroelastic problems
• Tensor networks combined with variational quantum circuits
• End-to-end trainable hybrid algorithm for classification/regression