Saved in:
| Main Authors: | , , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2510.07802 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Table of Contents:
- As quantum control approaches hardware-imposed performance limits, weak effects omitted by reduced models become consequential. Assumptions required for analytic tractability then cease to guide control design and instead constrain further improvement. Here, we relax such assumptions and use simulation-guided stochastic tree search to navigate combinatorially large, discrete pulse-sequence spaces for robust many-body spin control. Experimentally, in a solid-state spin ensemble, the resulting computationally discovered pulse sequences substantially outperform analytically optimized baselines, despite being excluded by construction from analytic design criteria. Importantly, these unconventional sequences expose predictive structural features that enable rapid neural network--based performance evaluation. This efficiency gain makes the combinatorial scaling tractable and expands the control alphabet from 8 symmetry-restricted pulses to over 26,000 hardware-resolved options. The resulting fine-grained design freedom provides the control resolution required to reliably address weak, performance-limiting effects, unlocking qualitatively different spin-control capabilities beyond decades of traditional sequence design. Together, these results show that near performance limits, simplifying assumptions can become a primary constraint on quantum control in realistic hardware, and must be repurposed to guide computational discovery.