Fully-connected VQCs reach 90–96% of quantum-transformer R² on tabular benchmarks with 40–50% fewer parameters, and expressibility saturates by circuit depth about three.
Quantum self-attention neural networks for text classification
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
quant-ph 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Do Quantum Transformers Help? A Systematic VQC Architecture Comparison on Tabular Benchmarks
Fully-connected VQCs reach 90–96% of quantum-transformer R² on tabular benchmarks with 40–50% fewer parameters, and expressibility saturates by circuit depth about three.