TBNQuantum
Fully quantum training of parametric circuits with Quantum Hamiltonian Descent
Publication: quant-ph updates on arXiv.org
Author: Marco Burato, Leonardo Banchi
Source published: 27/10/2026, 04:00:00 UTC
Briefing created: 27/10/2026, 04:19:17 UTC
Based on a feed excerpt; full article content has not been acquired. Source evidence retrieved 27/10/2026, 04:13:24 UTC.
Briefings
A new quantum algorithm called Quantum Hamiltonian Descent (QHD) trains parametric circuits by treating parameters as a quantum particle moving under a cost function potential. The study proves that for smooth periodic cost functions, simulation error bounds grow at most linearly with the number of parameters, unlike standard methods where error grows exponentially with qubits. Simulations on benchmark functions and parametric circuits demonstrate that optimized schedules significantly increase the probability of finding the global minimum compared to classical approaches.
Key points
- Quantum Hamiltonian Descent treats parameters as a quantum particle moving under a cost function potential to optimize circuits.
- For smooth periodic cost functions, simulation error bounds grow at most linearly with the number of parameters, avoiding exponential growth seen in standard methods.
Context
Parametric quantum circuits are typically trained using classical local optimizers that evaluate the cost function or gradient only at discrete points. This study proposes training these circuits entirely on the quantum computer using QHD, which controls circuit evolution via coherent qubit control on a second register.
Limitations
The study relies on simulations of benchmark functions and parametric circuits rather than experimental validation. The specific schedules that minimize gate count and maximize success probability remain theoretical; the exact numerical values of these schedules are not provided in the evidence.
Generated by Bittrees-hosted qwen35-2b sha256:aaf42c8b7c3c on 27/10/2026, 04:19:17 UTC. Based on publisher-provided feed evidence; may not cover the full original article. Check the source for details.
View archived briefing on IPFS
bafybeiaxakidzqicqvzycgt2te6s6cjfc2tvma26bvoyasxenum7rcy234