Classification of quantum-inspired machine learning relative to quantum machine learning

Determine whether quantum-inspired machine learning—defined as the use of concepts and tools from quantum mechanics and computing to design algorithms that run on classical hardware—should be classified within quantum machine learning or treated as a distinct, separate research area.

Background

The paper distinguishes quantum machine learning (QML), which uses quantum computers and parameterized quantum circuits for learning tasks, from quantum-inspired machine learning (QI-ML), which adapts ideas from quantum mechanics to classical algorithms without requiring quantum hardware. The authors note that QI-ML has been successfully applied to standard ML problems (e.g., graph time series, image classification) and quantum chemistry, and even used alongside QML on quantum hardware.

Despite these applications, the authors explicitly acknowledge an unresolved question regarding the research taxonomy: whether QI-ML should be considered part of QML or as a separate field. They emphasize that TorchQuantumDistributed provides tools suitable for developing both, indicating the practical relevance of clarifying this classification.