Quantum

Quantum-Classical Neural Networks for Solving Lindblad Master Equations

Read full article at source ↗

Based on a feed excerpt; full article content has not been acquired. Source evidence retrieved 19/10/2026, 04:13:24 UTC.

Briefings

A new variational framework combines classical neural networks with parameterized quantum circuits to solve Lindblad master equations via exact gradient evaluation. The method uses a purification on an enlarged Hilbert space to ensure physical properties of the density matrix, while comparing its covering number against purely quantum neural networks to establish a theoretical link between expressivity and the barren plateau phenomenon.

Key points

• The proposed method uses a compositional ansatz combining a classical neural network with a parameterized quantum circuit to minimize a residual cost function.
• Numerical experiments validate the accuracy of the method for representative Lindblad master equations, showing it outperforms purely quantum neural networks in gradient concentration.

Context

Variational quantum algorithms are promising for simulating open quantum systems but their expressivity and trainability are poorly understood. This work proposes a specific variational framework to address these limitations.

Limitations

The study relies on numerical experiments for validation and theoretical bounds for covering numbers. Specific experimental parameters and dataset sizes are not detailed in the provided evidence, limiting the reproducibility of the numerical results.

Generated by Bittrees-hosted qwen35-2b sha256:aaf42c8b7c3c on 19/10/2026, 04:33:03 UTC. Based on publisher-provided feed evidence; may not cover the full original article. Check the source for details.

View archived briefing on IPFS
bafybeif66regwk2shop43pswqntmxwua76kimaypynna3puquokpkcbddq