Validate theoretical trainability conditions on quantum hardware

Determine whether the theoretical trainability conditions derived for QPI-DeepONet-MAC translate into practical advantages when the architecture is implemented on quantum hardware.

Background

The paper establishes analytical bounds for the gradients of the quantum parameters in QPI-DeepONet-MAC and derives scaling conditions intended to mitigate barren-plateau effects. However, these results are theoretical and do not establish that the predicted trainability properties yield practical benefits on noisy intermediate-scale quantum devices.

The authors identify implementation and experimental validation on quantum hardware as future work, including the development of optimized circuit constructions, parameterizations, and training procedures adapted to hardware noise and resource constraints. The unresolved issue is whether the theoretically derived trainability conditions persist in such practical settings and lead to measurable advantages.

References

Such experiments will provide an opportunity to test whether the theoretical trainability conditions derived here translate into practical advantages on quantum hardware.

— QPI-DeepONet-MAC: A Scalable and Stable Hybrid Classical-Quantum Architecture for Physics-Informed Deep Operator Networks  (2610.01824 - Lantigua et al., 1 Oct 2026) in Section 5, Conclusions and future perspectives