Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Log inSign up
Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Blog
Job Board
Q3AS

© 2026 Aqora Quantum S.A.S.

TermsPrivacyLegal Notice

arXiv:2609.11248

Generative Replay Mitigates Sample Starvation in Quantum Architecture Search

Akash Kundu, Amit Kumar Jaiswal, Sebastian Feld, Prayag Tiwari

quant-phcs.AIcs.ETcs.LGSubmitted Sep 10, 2026• Announced Sep 11, 2026• 0 citations
arXivHTMLPDF

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.

Comments
GenQAS: 38 pages, 7 figures, 2 tables and 1 algorithm in main text

Related papers

TensorRL-QAS: Reinforcement learning with tensor networks for improved quantum architecture search

Akash Kundu, Stefano Mangini

Variational quantum algorithms hold the promise to address meaningful quantum problems already on noisy intermediate-scale quantum hardware.In spite of the promise, they face the challenge of designing quantum circuits that both solve the target problem and comply with device limitations.Quantum architecture search (QAS) automates the design process of quantum circuits, with reinforcement learning (RL) emerging as a promising approach.Yet, RL-based QAS methods encounter significant scalability issues, as computational and training costs grow rapidly with the number of qubits, circuit depth, and hardware noise.To address these challenges, we introduce TensorRL-QAS, an improved framework that combines tensor network methods with RL for QAS.By warm-starting the QAS with a matrix product state approximation of the target solution, TensorRL-QAS effectively narrows the search space to physically meaningful circuits and accelerates the convergence to the desired solution.Tested on several quantum chemistry problems of up to 12-qubit, TensorRL-QAS achieves up to a 10-fold reduction in CNOT count and circuit depth compared to baseline methods, while maintaining or surpassing chemical accuracy.It reduces classical optimizer function evaluation by up to 100-fold, accelerates training episodes by up to 98%, and can achieve 50% success probability for 10-qubit systems, far exceeding the <1% rates of baseline.Robustness and versatility are demonstrated both in the noiseless and noisy scenarios, where we report a simulation of an 8-qubit system.Furthermore, TensorRL-QAS demonstrates effectiveness on systems on 20-qubit quantum systems, positioning it as a state-of-the-art quantum circuit discovery framework for near-term hardware and beyond.

arXiv:2505.09371•neural information processing systems•Jan 1, 2025•8 citations

QAS-QTNs: Curriculum Reinforcement Learning-Driven Quantum Architecture Search for Quantum Tensor Networks

Siddhant Dutta, Nouhaila Innan, Sadok Ben Yahia +1 more

Quantum Architecture Search (QAS) is an emerging field aimed at automating the design of quantum circuits for optimal performance. This paper introduces a novel QAS framework employing hybrid quantum reinforcement learning with quantum curriculum learning strategies, enabling learning agents to tackle increasingly complex quantum circuit design tasks. We benchmark four state-of-the-art classical reinforcement learning algorithms (A2C, PPO, DDQN, TD3) against their quantum-enhanced counterparts (QA2C, QPPO, QDDQN, QTD3) for optimizing variational quantum circuits (VQCs). Our approach progressively increases circuit depth and gate complexity during training, leveraging parameterized quantum circuits as function approximations. To improve learning efficiency and stability, all algorithms, both classical and quantum, are augmented with Prioritized Experience Replay (PER). Experimental results show that quantum-enhanced RL significantly outperforms classical methods. In a 2-qubit environment, PERQDDQN achieves a success probability of 0.46 with 3,000 optimal successes, surpassing classical PERDDQN ($\mathbf{0. 4 2}, \mathbf{2, 4 0 0}$). In the more complex 3-qubit setting, PERQDDQN and PERQTD3 reach success probabilities of 0.47, with optimal success counts of 3,800 and 3,600, respectively, outperforming their classical counterparts. Additionally, we apply our QAS-QTN approach to a classification problem, where the optimized quantum circuit achieves an accuracy of 90.33%, outperforming quantum models consisting of random ansatz. This hybrid classical-quantum approach leads to faster convergence and more efficient quantum circuit designs, demonstrating its potential for advancing automated quantum architecture search.

arXiv:2507.12013•IEEE International Conference on Quantum Computing and Engineering (QCE)•Aug 30, 2025•8 citations

DreamQAS: Learning a Decision-Useful World Model for VQE-Efficient Quantum Architecture Search

Jiayang Niu, Yan Wang, Jie Li +4 more

Reinforcement-learning-based quantum architecture search (RL-QAS) repeatedly invokes a variational quantum eigensolver (VQE) after each gate addition even though circuit transitions and action legality are known. DreamQAS preserves these exact dynamics and learns only expensive post-VQE feedback through a recurrent ensemble that predicts a frontier-relative feedback score without requiring the exact ground-state energy, enabling uncertainty-controlled multi-step imagination. Under a common 15,000-episode budget and frozen evaluation, DreamQAS has the lowest reported mean error among RL methods on all five main molecular tasks. At fine-error targets reached by all seeds of DreamQAS and a matched non-imaginative control, it uses 1.6-2.0 times fewer real VQE calls on four tasks. Holding LiH-4q feedback-model weights fixed, its imagined-policy actor attains 0.073 mHa, versus 4.280 mHa and 4.434 mHa for greedy and beam deployment. Learned-transition and end-to-end predictor controls further show that preserving exact circuit structure and using feedback through policy learning are both important. Counterfactual action-ranking improves throughout training on all five probed tasks, while ensemble disagreement improves risk-coverage over random rejection on three tasks. DreamQAS therefore learns decision-useful feedback for QAS without modeling already-known circuit dynamics or requiring the exact ground-state energy.

arXiv:2607.29491•Jul 31, 2026•0 citations

Replay-buffer engineering for noise-aware quantum circuit optimization

Akash Kundu, Sebastian Feld

Deep reinforcement learning for quantum circuit optimization faces three bottlenecks: replay buffers that overlook temporal difference (TD) target reliability, curriculum-based architecture search requiring a full quantum-classical evaluation after every edit, and the discard of noiseless trajectories when retraining under hardware noise. We address these limitations by treating replay as a central algorithmic lever. We introduce ReaPER+, an annealed replay rule that transitions from TD-error prioritization to reliability-aware sampling as value estimates mature. ReaPER+ achieves up to 4x higher sample efficiency than fixed PER, ReaPER, and uniform replay, while matching prior on-policy solution quality with up to 32x fewer interactions At 12 qubits, fixed ReaPER reaches the lowest energy error in the fewest steps, while PER and uniform replay find more compact circuits at higher error. On tasks scaling to 20 qubits, ReaPER+ retains its advantage, demonstrating that reliability-aware annealing extends beyond small-system benchmarks. LunarLander-v3 confirms that the ReaPER+ is domain-agnostic, it improves success rates by up to 26.8% over PER and 21.8% over fixed ReaPER, with a 3% AUC gain over both. We further introduce OptCRLQAS, which amortizes quantum-classical evaluations across multiple architectural edits, reducing training wall-clock time by up to 67.5% on 12-qubit without degrading solution quality. Finally, lightweight replay-buffer transfer warm-starts noisy optimization from noiseless trajectories, without weight transfer or $ε$-greedy pretraining, reducing steps to chemical accuracy by 85-90% and final energy error by up to 90% relative to from-scratch learning. Transfer gains increase with system size. Together, these results establish experience storage, sampling, and transfer as decisive levers for sample efficient, noise-aware quantum circuit optimization.

quant-pharXiv:2604.21863•Apr 24, 2026•1 citation

General-purpose quantum architecture search based on deep reinforcement learning

Xiao-Yu Bi, Y. Yu, Ye‐Hong Chen +1 more

Reinforcement learning (RL) shows promise for automated quantum circuit design but often stalls due to a “fidelity trap”: by optimizing only state fidelity, agents overlook entanglement structure and become stranded in suboptimal, overly complex circuits, resulting in significantly reduced search efficiency. In this paper, we propose a scheme that overcomes this barrier by implementing an entanglement-aware learning framework and enhancing the agent's reward function with a direct, quantitative measure of entanglement. This approach offers a more comprehensive physical description of the state space. We demonstrate the efficacy of this principle on three- and four-qubit state-synthesis tasks within an expanded gate set. For this problem, where the fidelity-driven agent systematically fails to discover the minimal-depth circuit, our entanglement-aware agent consistently succeeds. This transformative result is highly robust against variations in initial random seeds and extends to multiqubit systems even in the presence of noise. Our findings establish a generalizable principle that incorporating entanglement as an auxiliary reward can significantly enhance RL-based solutions for a broad class of fidelity-centric tasks in quantum physics and pave the way for scalable, automated discovery on near-term quantum devices.

Physical Review A•Sep 26, 2025•10 citations

Distributed quantum architecture search using multi-agent reinforcement learning

Mikhail Sergeev, Georgii Paradezhenko, Daniil Rabinovich +1 more

Quantum architecture search (QAS) automates the design of parameterized quantum circuits for variational quantum algorithms. The framework finds a well-suited problem-specific structure of a variational ansatz. Among possible implementations of QAS the reinforcement learning (RL) stands out as one of the most promising. Current RL approaches are single-agent-based and show poor scalability with a number of qubits due to the increase of the action space dimension and the computational cost. We propose a novel multi-agent RL algorithm for QAS with each agent acting separately on its own block of a quantum circuit. This procedure allows to significantly accelerate the convergence of the RL-based QAS and reduce its computational cost. We benchmark the proposed algorithm on MaxCut problem on 3-regular graphs and on ground energy estimation for the Schwinger Hamiltonian. In addition, the proposed multi-agent approach naturally fits into the set-up of distributed quantum computing, favoring its implementation on modern intermediate scale quantum devices.

quant-pharXiv:2511.22708•Dec 1, 2025•0 citations
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 7.0×7.0\times7.0× over passive replay. On a 15-qubit transverse field Ising model, GenQAS increases success probability from 12%12\%12% to 21%21\%21%. In a noisy 6-qubit BeH2_22​ transfer experiment, generative replay reduces the steps to chemical accuracy by 92.7%92.7\%92.7%. These results show that generative replay can mitigate sample starvation in quantum architecture search and support more resource efficient circuit discovery.
Variational quantum algorithms hold the promise to address meaningful quantum problems already on noisy intermediate-scale quantum hardware.In spite of the promise, they face the challenge of designing quantum circuits that both solve the target problem and comply with device limitations.Quantum architecture search (QAS) automates the design process of quantum circuits, with reinforcement learning (RL) emerging as a promising approach.Yet, RL-based QAS methods encounter significant scalability issues, as computational and training costs grow rapidly with the number of qubits, circuit depth, and hardware noise.To address these challenges, we introduce TensorRL-QAS, an improved framework that combines tensor network methods with RL for QAS.By warm-starting the QAS with a matrix product state approximation of the target solution, TensorRL-QAS effectively narrows the search space to physically meaningful circuits and accelerates the convergence to the desired solution.Tested on several quantum chemistry problems of up to 12-qubit, TensorRL-QAS achieves up to a 10-fold reduction in CNOT count and circuit depth compared to baseline methods, while maintaining or surpassing chemical accuracy.It reduces classical optimizer function evaluation by up to 100-fold, accelerates training episodes by up to 98%, and can achieve 50% success probability for 10-qubit systems, far exceeding the <1% rates of baseline.Robustness and versatility are demonstrated both in the noiseless and noisy scenarios, where we report a simulation of an 8-qubit system.Furthermore, TensorRL-QAS demonstrates effectiveness on systems on 20-qubit quantum systems, positioning it as a state-of-the-art quantum circuit discovery framework for near-term hardware and beyond.
Quantum Architecture Search (QAS) is an emerging field aimed at automating the design of quantum circuits for optimal performance. This paper introduces a novel QAS framework employing hybrid quantum reinforcement learning with quantum curriculum learning strategies, enabling learning agents to tackle increasingly complex quantum circuit design tasks. We benchmark four state-of-the-art classical reinforcement learning algorithms (A2C, PPO, DDQN, TD3) against their quantum-enhanced counterparts (QA2C, QPPO, QDDQN, QTD3) for optimizing variational quantum circuits (VQCs). Our approach progressively increases circuit depth and gate complexity during training, leveraging parameterized quantum circuits as function approximations. To improve learning efficiency and stability, all algorithms, both classical and quantum, are augmented with Prioritized Experience Replay (PER). Experimental results show that quantum-enhanced RL significantly outperforms classical methods. In a 2-qubit environment, PERQDDQN achieves a success probability of 0.46 with 3,000 optimal successes, surpassing classical PERDDQN (0.42,2,400\mathbf{0. 4 2}, \mathbf{2, 4 0 0}0.42,2,400). In the more complex 3-qubit setting, PERQDDQN and PERQTD3 reach success probabilities of 0.47, with optimal success counts of 3,800 and 3,600, respectively, outperforming their classical counterparts. Additionally, we apply our QAS-QTN approach to a classification problem, where the optimized quantum circuit achieves an accuracy of 90.33%, outperforming quantum models consisting of random ansatz. This hybrid classical-quantum approach leads to faster convergence and more efficient quantum circuit designs, demonstrating its potential for advancing automated quantum architecture search.
Reinforcement-learning-based quantum architecture search (RL-QAS) repeatedly invokes a variational quantum eigensolver (VQE) after each gate addition even though circuit transitions and action legality are known. DreamQAS preserves these exact dynamics and learns only expensive post-VQE feedback through a recurrent ensemble that predicts a frontier-relative feedback score without requiring the exact ground-state energy, enabling uncertainty-controlled multi-step imagination. Under a common 15,000-episode budget and frozen evaluation, DreamQAS has the lowest reported mean error among RL methods on all five main molecular tasks. At fine-error targets reached by all seeds of DreamQAS and a matched non-imaginative control, it uses 1.6-2.0 times fewer real VQE calls on four tasks. Holding LiH-4q feedback-model weights fixed, its imagined-policy actor attains 0.073 mHa, versus 4.280 mHa and 4.434 mHa for greedy and beam deployment. Learned-transition and end-to-end predictor controls further show that preserving exact circuit structure and using feedback through policy learning are both important. Counterfactual action-ranking improves throughout training on all five probed tasks, while ensemble disagreement improves risk-coverage over random rejection on three tasks. DreamQAS therefore learns decision-useful feedback for QAS without modeling already-known circuit dynamics or requiring the exact ground-state energy.
Deep reinforcement learning for quantum circuit optimization faces three bottlenecks: replay buffers that overlook temporal difference (TD) target reliability, curriculum-based architecture search requiring a full quantum-classical evaluation after every edit, and the discard of noiseless trajectories when retraining under hardware noise. We address these limitations by treating replay as a central algorithmic lever. We introduce ReaPER+, an annealed replay rule that transitions from TD-error prioritization to reliability-aware sampling as value estimates mature. ReaPER+ achieves up to 4x higher sample efficiency than fixed PER, ReaPER, and uniform replay, while matching prior on-policy solution quality with up to 32x fewer interactions At 12 qubits, fixed ReaPER reaches the lowest energy error in the fewest steps, while PER and uniform replay find more compact circuits at higher error. On tasks scaling to 20 qubits, ReaPER+ retains its advantage, demonstrating that reliability-aware annealing extends beyond small-system benchmarks. LunarLander-v3 confirms that the ReaPER+ is domain-agnostic, it improves success rates by up to 26.8% over PER and 21.8% over fixed ReaPER, with a 3% AUC gain over both. We further introduce OptCRLQAS, which amortizes quantum-classical evaluations across multiple architectural edits, reducing training wall-clock time by up to 67.5% on 12-qubit without degrading solution quality. Finally, lightweight replay-buffer transfer warm-starts noisy optimization from noiseless trajectories, without weight transfer or εεε-greedy pretraining, reducing steps to chemical accuracy by 85-90% and final energy error by up to 90% relative to from-scratch learning. Transfer gains increase with system size. Together, these results establish experience storage, sampling, and transfer as decisive levers for sample efficient, noise-aware quantum circuit optimization.
Reinforcement learning (RL) shows promise for automated quantum circuit design but often stalls due to a “fidelity trap”: by optimizing only state fidelity, agents overlook entanglement structure and become stranded in suboptimal, overly complex circuits, resulting in significantly reduced search efficiency. In this paper, we propose a scheme that overcomes this barrier by implementing an entanglement-aware learning framework and enhancing the agent's reward function with a direct, quantitative measure of entanglement. This approach offers a more comprehensive physical description of the state space. We demonstrate the efficacy of this principle on three- and four-qubit state-synthesis tasks within an expanded gate set. For this problem, where the fidelity-driven agent systematically fails to discover the minimal-depth circuit, our entanglement-aware agent consistently succeeds. This transformative result is highly robust against variations in initial random seeds and extends to multiqubit systems even in the presence of noise. Our findings establish a generalizable principle that incorporating entanglement as an auxiliary reward can significantly enhance RL-based solutions for a broad class of fidelity-centric tasks in quantum physics and pave the way for scalable, automated discovery on near-term quantum devices.
Quantum architecture search (QAS) automates the design of parameterized quantum circuits for variational quantum algorithms. The framework finds a well-suited problem-specific structure of a variational ansatz. Among possible implementations of QAS the reinforcement learning (RL) stands out as one of the most promising. Current RL approaches are single-agent-based and show poor scalability with a number of qubits due to the increase of the action space dimension and the computational cost. We propose a novel multi-agent RL algorithm for QAS with each agent acting separately on its own block of a quantum circuit. This procedure allows to significantly accelerate the convergence of the RL-based QAS and reduce its computational cost. We benchmark the proposed algorithm on MaxCut problem on 3-regular graphs and on ground energy estimation for the Schwinger Hamiltonian. In addition, the proposed multi-agent approach naturally fits into the set-up of distributed quantum computing, favoring its implementation on modern intermediate scale quantum devices.