RubriQ is introduced, a scalable framework that formulates circuit synthesis as a large language model (LLM) code-generation task, optimized via group relative policy optimization (GRPO), which establishes an automated, high-performance computing (HPC)-driven pipeline for generating hardware-ready, fault-tolerant quantum circuits at scale.
Abstract
Designing fault-tolerant quantum circuits that are both algorithmically correct and hardware compatible remains a major bottleneck in the transition to scalable quantum computing. We introduce RubriQ, a scalable framework that formulates circuit synthesis as a large language model (LLM) code-generation task, optimized via group relative policy optimization (GRPO). Unlike conventional black-box neural critics, RubriQ employs a domain-grounded programmatic rubric as the reinforcement learning reward function, evaluating circuits for T-gate reduction, hardware topology compliance, and unitary fidelity. To support high-throughput training, RubriQ integrates GPU-accelerated CUDA-Q simulation directly into the reinforcement learning (RL) loop and is deployed on NERSC Perlmutter using DeepSpeed ZeRO2 across multinode NVIDIA A100 clusters. On benchmark tasks, RubriQ achieves a mean T-gate compression of 3.31x, significantly outperforming sparse-reward RL baselines (2.05x), converging 2-3x faster, and maintaining less than 1\% hardware-constraint violations. Validated on IBM and IonQ quantum processors, RubriQ establishes an automated, high-performance computing (HPC)-driven pipeline for generating hardware-ready, fault-tolerant quantum circuits at scale.
Distributed quantum computing provides a scalable route for executing quantum circuits beyond the capacity limits of a single quantum processing unit (QPU), but it introduces a communication-aware compilation problem involving strict hardware constraints and circuit dependencies. This paper presents an architecture-aware reinforcement-learning framework that formulates distributed quantum compilation as a constrained Markov Decision Process (MDP). The compiler-level communication actions dynamically update logical-qubit placement and enable subsequent gate execution. A heterogeneous graph model represents interactions among hardware, logical qubits, and circuit operations, while a policy trained via Proximal Policy Optimization optimizes EPR-pair consumption and communication makespan. Evaluation across benchmark circuits shows that our policy matches state-of-the-art heuristics on structured workloads, with lookahead reward shaping yielding modest improvements on unstructured circuits. These results demonstrate that reinforcement learning is a flexible alternative to manual heuristics, though scalability remains a key bottleneck for practical use.
Chien-Tung Kuo, Felix Burt, Samuel Yen-Chi Chen et al.· 0 citations
A reinforcement learning framework that embeds a deterministic Commutation-and-Reduction (CR) algorithm directly into the training environment, enabling the agent to focus its learning capacity on the non-trivial optimizations where reinforcement learning adds real value.
Khoa Dang Tao, Sumin Jin, M. Raza et al.· 0 citations
Reinforcement-learning-based quantum architecture search (RL-QAS) repeatedly optimizes a variational quantum eigensolver (VQE) after extending a circuit, although circuit construction and action legality are deterministic and known. We introduce DreamQAS, a model-based RL framework that preserves these exact circuit dynamics and learns only the expensive post-VQE feedback. A recurrent randomized-prior ensemble predicts an oracle-free score relative to an empirical energy frontier and supports multi-step imagined policy learning over explicit legal circuits. Ranking-based activation, uncertainty-aware pessimism and truncation, and selective real-VQE verification form a reliability-controlled learning loop. Under a common 15,000-episode budget and frozen evaluation for the RL methods, DreamQAS has the lowest mean frozen-policy energy error on four of five molecular tasks and the second-lowest on one. At fine-error targets reached by all seeds of both methods, it uses 1.6x to 2.0x fewer real VQE calls on four tasks and 10.6x fewer on BeH2-8q. Counterfactual action-ranking utility increases across all five tasks, with a mean increase of 0.346 and a 95 percent confidence interval of [0.185, 0.507], while direct greedy and beam use of the same model does not recover the gains of imagined policy learning. Ensemble disagreement also improves risk-coverage over random rejection on all three probed tasks. These results establish a world-model design for QAS whose value lies in decision-useful feedback rather than exact energy prediction.
The contrast between settings is the central finding: when approximate outputs can be rescued by post-processing, the transformer succeeds; when exact discrete correctness is required, autoregressive drift limits reliability, with both inference-time search and data scaling as effective levers while training-side fine-tuning and model-level diversification are not.
Despite rapid progress in quantum optimization, the field lacks broad real-hardware benchmarks comparing multiple algorithmic families across diverse classically hard combinatorial problems under one protocol. We present a hardware-aware benchmark of gate-based quantum optimization across four NP-hard problems: the multi-dimensional knapsack problem (MDKP), maximum independent set (MIS), quadratic assignment problem (QAP), and market-share problem, spanning variational methods (VQE, CVaR-VQE), standard, multi-angle, and warm-start QAOA, and qubit-efficient encodings (Pauli correlation encoding (PCE), QRAO), executed on IBM Heron r1/r2 processors under resilience-level-2 mitigation. To our knowledge this includes the first real-hardware QRAO results and the first multi-problem PCE hardware benchmark. Across 247 method–instance combinations we report transpiled circuit size, an independent-error gate-count fidelity proxy Fest, and hardware outcomes. An empirical operating point near Fest≈0.1 ( ∼770 two-qubit gates at the median Heron-r2 CZ error rate) marks the transition to noise-dominated execution in the MDKP and MIS regimes. Two limitations emerge. QAP couples dense one-hot encodings with an exponentially sparse feasible manifold (feasible fraction 10!/2100 at n=10); no tested hardware method returns a feasible assignment. The tested QAOA-family circuits become noise-dominated after compilation, and a matched uniform-random control shows most feasible low-fidelity outcomes lie within the random range, with one MIS warm-start result reported as a finite-sample exception. A compilation counterfactual (SWAP-aware, fractional-gate, Nighthawk-topology) reduces two-qubit counts but moves no circuit above Fest=10−3; conclusions therefore apply to the tested implementations, not QAOA in general. Qubit-efficient methods extend runnable instance sizes but gain only within the empirical fidelity budget33 Code available at: https://github.com/SMU-Quantum/quantum-optimization-benchmarks.. Code available at: https://github.com/SMU-Quantum/quantum-optimization-benchmarks.
Monit Sharma, HoongChuin Lau· Quantum Science and Technolo...· 1 citation
: Quantum logic reversible synthesis is a fundamental operation in quantum computing. One of the most challenging issues in this field resides in navigating the immense search space to synthesize the most compact circuit configurations, which are critical for realizing reliable, noise-free, and error-free quantum computing systems. To address this challenge, this study proposes a novel hypercube-encoded quantum-inspired optimization framework to formulate the synthesis task as a trajectory-finding process. This structure-informed domain knowledge transformation delivers exceptional search direction guidance, moving away from blind, black-box exploration. Specifically, by mapping the reversible functions onto the hypercube architecture, the framework embeds explicit dual Hamming-distance (HMD) guidance metrics into a global-best guided quantum-inspired tabu search (GQTS) engine. To minimize computational cost and enhance search efficiency, the framework incorporates a domain-informed initialization and couples a streamlined two-particle configuration with a global-best mechanism, thereby amplifying the efficiency of the underlying quantum-inspired updating mechanism to escape local optima and rapidly converge once a successfully synthesized superior path is locked. Under an online step relaxation mechanism, the framework preserves exceptional structural optimization flexibility without altering the underlying hypercube representation. The framework’s significance is rigorously evaluated against both rigid hypercube rule-based methods and generic randomized search-based heuristics, using gate count and exact-optimality as primary evaluation metrics. Extensive ablation studies first validate the individual and synergistic contributions of each core algorithmic component. Experimental results demonstrate that for the complete set of 3-bit reversible functions, the proposed HMD-guided GQTS (HMD-GQTS) framework achieves over 98% exact-optimal circuits in a single fine-tuned sweep, with selective retries attaining 100% exhaustive optimal coverage. Furthermore, for typical 4-bit benchmark instances, the method consistently delivers competitive gate counts, matching or improving upon previous hypercube-based outcomes. Through the seamless integration of structure-informed guidance and coordinated optimization mechanisms, the framework preserves exceptional structural flexibility and algorithmic robustness, offering an effective, low-cost, and highly scalable avenue for reversible circuit synthesis.
Yu-Chi Jiang· Computers, Materials & C...· 0 citations