Quantum circuit search improves with synthetic replay model
Generative Replay Mitigates Sample Starvation in Quantum Architecture Search
Artificial IntelligenceEmerging TechnologiesMachine Learning
Summary
Finding useful designs for quantum circuits can be very hard because the search space is huge and good solutions are rare. The authors introduce GenQAS, a method that teaches a computer to imagine new possible steps in building quantum circuits based on real examples it has seen. By mixing these imagined steps with real experiences, the learning process becomes more efficient. Testing this approach on chemical and physics problems showed it finds better circuits faster and more reliably. This helps make quantum circuit design more practical and resource efficient.
What this means in practice
- •For quantum algorithm developers: Design quantum circuits more efficiently by generating synthetic training data to guide reinforcement learning models.
- •For computational chemists: Accelerate quantum simulations of molecules by improving the search for compact circuits that approximate molecular energies accurately.
Authors
Akash Kundu, Amit Kumar Jaiswal, Sebastian Feld, Prayag Tiwari
Abstract
Reinforcement learning (RL) can automate quantum architecture search, but its scalability is limited when useful circuit trajectories become rare in the rapidly expanding search space. Existing replay mechanisms reuse observed transitions; the proposed learned model produces additional predicted one step transitions from real state-action seeds. Here we introduce GenQAS, a tensor network-guided RL framework that combines a fixed matrix product state warm-start with prioritized generative replay. A learned local transition model generates synthetic circuit transitions on demand and mixes them with real experience during Double Deep Q-Network updates. Under a random exploration analysis, near ground state circuits occupy a rapidly shrinking region of the accessible state space. We investigate whether real data anchored synthetic replay can improve the effective training signal in this regime. Across chemical Hamiltonian benchmarks from 6 to 12 qubits, GenQAS improves fixed-budget success probability and identifies compact circuits at competitive energy error. At 12 qubits, it improves final success probability by up to $7.0\times$ over passive replay. On a 15-qubit transverse field Ising model, GenQAS increases success probability from $12\%$ to $21\%$. In a noisy 6-qubit BeH$_2$ transfer experiment, generative replay reduces the steps to chemical accuracy by $92.7\%$. These results show that generative replay can mitigate sample starvation in quantum architecture search and support more resource efficient circuit discovery.