Hybrid Quantum-Classical Mixture of Experts: Unlocking Topological Advantage via Interference-Based Routing
Reda HEDDAD, Dr. Lamiae Bouanane · 2025
Read Hybrid Quantum-Classical Mixture of Experts: Unlocking Topological Advantage via Interference-Based Routing by Reda HEDDAD, Dr. Lamiae Bouanane (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2512-22296.txt.
Description
Reda Heddad and Lamiae Bouanane propose QMoE, a mixture-of-experts network whose router is a small quantum circuit. Mixture-of-experts models scale by sending each input to a few specialist sub-networks, but classic routers struggle with tangled decision boundaries and uneven expert use. The authors test whether quantum interference can route better, isolating the effect through an ablation study.
Their quantum gating network uses angle embedding and wave interference, effectively acting as a high-dimensional kernel method. On non-linearly separable data such as the Two Moons dataset, it separated distributions that linear classic routers handle poorly, with better parameter efficiency. Simulations also checked robustness against quantum noise, supporting feasibility on near-term hardware. Federated and privacy-preserving learning are among the suggested applications.
Catalog record
| CREATOR | Reda HEDDAD, Dr. Lamiae Bouanane |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 27 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2512-22296