AI and the World's Hard Problems · Quantum physics & quantum computing · arXiv · 2026-10-26
HQTN-SER: Speech Emotion Recognition with Hybrid Quantum Tensor Networks
Speech emotion recognition remains fragile in real-world conditions because emotional cues are subtle and speaker-dependent. The authors propose HQTN-SER, a hybrid quantum-classical model that fuses an MPS-inspired quantum tensor network, using three or four qubits, with a classical latent embedding. It reached 80.09% accuracy on RAVDESS, 78.12% on SAVEE, and 73.51% on MDER, and training converged stably.