Chapter 4 of dossier QC-Accelerate-002: 'The loop eats first.' The dossier's thesis is a propose-refute-simulate-score-learn loop; this chapter runs the loop's cheapest case -- a search for quantum error-correcting codes, where 'making' a candidate means simulating it. It is, to the author's knowledge, the first AI-driven code search refereed at circuit level (finite-bias, faulty syndrome extraction) rather than at the flattering code-capacity grade that every published AI code search uses. The referee is a two-mode evaluator anchored to Amazon's published Ocelot cat-qubit chip (three independently measured closures with no fitting) and corridor-validated against an independent published simulation. Four priors were pre-registered and cryptographically timestamped before the search ran; two resolved true, two false. The measured, model-level payload: codes must match their bias regime. A searched [24,4,12] code (three-fold repetition crossed with the [8,4,4] extended Hamming code) beats budget-matched repetition by about four times at the future high-bias operating point, and inverts to 12-63 times worse when the gates are made ten times dirtier -- so the advantage exists only in the regime the search was aimed at. This is a proof of principle: the loop turned once, over a small hand-seeded candidate set, with scoring automated and proposals human/LLM-guided -- not an exhausted search, and explicitly not a hardware demonstration. Ships with the explicit winning code, an eleven-source citation audit, a red-team led by the model-versus-hardware caveat, a ten-entry ledger of the machinery's own caught-and-published errors, and a reproducibility section (noise model, decoder configuration, committed scripts, pinned software stack, and the Ocelot data DOI). An Open Dossier publication: executable claims under CI, published adversarial review, OpenTimestamps-anchored releases.
Chapter 4 of dossier QC-Accelerate-002: 'The loop eats first.' The dossier's thesis is a propose-refute-simulate-score-learn loop; this chapter runs the loop's cheapest case -- a search for quantum error-correcting codes, where 'making' a candidate means simulating it. It is, to the author's knowledge, the first AI-driven code search refereed at circuit level (finite-bias, faulty syndrome extraction) rather than at the flattering code-capacity grade that every published AI code search uses. The referee is a two-mode evaluator anchored to Amazon's published Ocelot cat-qubit chip (three independently measured closures with no fitting) and corridor-validated against an independent published simulation. Four priors were pre-registered and cryptographically timestamped before the search ran; two resolved true, two false. The measured, model-level payload: codes must match their bias regime. A searched [24,4,12] code (three-fold repetition crossed with the [8,4,4] extended Hamming code) beats budget-matched repetition by about four times at the future high-bias operating point, and inverts to 12-63 times worse when the gates are made ten times dirtier -- so the advantage exists only in the regime the search was aimed at. This is a proof of principle: the loop turned once, over a small hand-seeded candidate set, with scoring automated and proposals human/LLM-guided -- not an exhausted search, and explicitly not a hardware demonstration. Ships with the explicit winning code, an eleven-source citation audit, a red-team led by the model-versus-hardware caveat, a ten-entry ledger of the machinery's own caught-and-published errors, and a reproducibility section (noise model, decoder configuration, committed scripts, pinned software stack, and the Ocelot data DOI). An Open Dossier publication: executable claims under CI, published adversarial review, OpenTimestamps-anchored releases.
Quantum error correction (QEC) is indispensable for scalable fault-tolerant quantum computing. However, discovering QEC codes that remain effective is challenging, as logical performance depends on the interplay between code structure, hardware, syndrome extraction, and decoding, which often impose competing requirements. Here we introduce OmniQEC, an efficient AI scientist for discovering QEC codes suited to deployment on modern quantum processors. OmniQEC formulates QEC design as an iterative discovery process in which an orchestrator, implemented by advanced large language models (LLMs), coordinates code generation, code-level screening, syndrome-extraction synthesis, and decoder-based circuit evaluation. At its core, OmniQEC combines a self-evolving reasoning mechanism with a slow--fast synergistic workflow: a fast loop explores candidates using inexpensive code-level proxies, whereas a slow loop performs physically grounded circuit-level evaluation and feeds the resulting evidence back into the search. We evaluate OmniQEC across four qLDPC construction families, three LLM backends, and 14 total-physical-qubit budgets per backend. The discovered codes show steadily improving logical-error suppression with increasing physical-qubit budgets and outperform the BB codes with [[72,12,6]] and [[144,12,12]] under complete-implementation budgets of 98 and 240 physical qubits, respectively. The discovered codes are hardware-friendly and may be of independent interest for practical QEC implementation. These findings pave the way towards LLM-assisted QEC discovery grounded in physically informed code--circuit--decoder co-design.
Compiling quantum algorithms for real devices involves hard optimization problems that can decide whether an algorithm runs at scale. We present a general method for solving such problems with AI agents, with every result certified. It builds on autoresearch loops, in which agents iteratively improve programs against a fixed evaluator and which have recently improved quantum cryptanalysis circuits, and adds two elements: a frozen, exact verifier certifies every result, and this verifier, together with the challenge around it, is built by an agent that formulates the problem with a researcher and validates the verifier against a solved case, deliberately broken solutions and published results before any solution exists. The method applies to any optimization problem whose solutions can be verified exactly and efficiently. We apply it to three compilation problems in quantum simulation, namely fermion-to-qubit encoding, Hamiltonian layout and unitary synthesis. In each, the agents improved substantially on published or standard constructions, reducing the cost being optimized by up to 73% at the largest sizes tested and outperforming a state-of-the-art optimizer for fermion-to-qubit encodings. Every submission is a program accompanied by its agent's notes, so the ideas behind each improvement are recorded in human-readable form; later agents built on them, and we describe them here. These results open the way to automated, certified discovery across quantum circuit design and beyond.
The integration of quantum computing and artificial intelligence (AI) constitutes a bidirectional synergy that is reshaping both disciplines. This review investigates the reciprocal relationship in which AI addresses foundational challenges in quantum computing, while quantum computing offers the potential to advance machine learning beyond classical constraints. Recent advancements exemplify this interaction: Google DeepMind’s AlphaQubit neural network decoder has achieved state-of-the-art quantum error correction, improving performance by 6% over tensor networks and 30% over correlated matching methods. Additionally, quantum neural networks have demonstrated exponential improvements in sample complexity for specific learning tasks. This review systematically examines the current landscape of quantum-AI integration across three primary dimensions. First, it addresses AI-enhanced quantum systems, such as transformer-based error correction, reinforcement learning for circuit optimization, and AI-driven hardware calibration. Second, it explores quantum-accelerated machine learning algorithms, including variational quantum neural networks, quantum generative adversarial networks, and quantum reinforcement learning. Third, it evaluates industry deployments in sectors such as life sciences, financial services, climate modeling, and pharmaceutical development. Notable examples include IonQ’s quantum chemistry simulations, which achieved 40% efficiency improvements in carbon capture material design; St. Jude’s identification of KRAS protein inhibitors using quantum machine learning with experimental validation; and JP Morgan Chase’s implementation of quantum portfolio optimization. The analysis indicates that the field has advanced from preliminary demonstrations to production-grade applications, especially in drug discovery and molecular simulation, where quantum computing offers measurable benefits. Nevertheless, several challenges remain, including barren plateaus in variational algorithms, scalability constraints in current noisy intermediate-scale quantum (NISQ) devices, requirements for real-time decoding speed, and the necessity for fault-tolerant quantum computing (FTQC) systems. IBM’s 2029 roadmap, which targets 200 logical qubits capable of executing 100 million gates, together with progress in high-rate quantum low-density parity-check codes, outlines a trajectory toward practical fault tolerance. This review makes two principal research contributions: Comprehensive Evaluation Framework for Quantum Advantage Assessment: We establish the first systematic, integrated methodology for determining quantum computing feasibility that combines problem characterization, resource estimation, quantum advantage assessment, and quantum algorithm paradigm selection. This framework consolidates criteria scattered across academic literature and industry practice into a unified decision-making tool applicable across chemistry, optimization, machine learning, and simulation domains. The framework addresses a critical industry need: enabling non-expert practitioners (chemists, financial analysts, materials scientists) to objectively assess quantum computing suitability without requiring deep quantum expertise. Novel Quantum Resource Optimization Algorithms: We present three concrete algorithmic contributions advancing quantum-AI integration: • Data Encoding Efficiency Algorithm that automatically selects optimal qubit encoding strategies (amplitude vs. angle encoding) minimizing total quantum resource consumption • Error Budget Optimization Algorithm that iteratively determines optimal quantum error correction code distance balancing logical error rate targets against physical qubit overhead • Real-time Hardware Specification Aggregation from multiple quantum cloud platforms (IBM Quantum, Amazon Braket, Google, IonQ) enabling dynamic feasibility assessment as hardware capabilities evolve Combined, these contributions establish quantum computing advantage assessment as a rigorous, data-driven discipline rather than ad-hoc expert judgment. The framework enables strategic quantum computing investment decisions, accelerates problem identification for early quantum utility, and provides a reference methodology for standardizing quantum advantage evaluation across academia and industry. This review synthesizes insights from recent academic literature, industry implementations, and expert perspectives to deliver a comprehensive assessment of quantum-AI synergies grounded in this evaluation framework. The mathematical foundations, ranging from quantum Fourier transforms to Gaussian processes, are examined. Best practices for researchers and practitioners are also outlined. The findings suggest that, although universal quantum advantage has not yet been realized, domain-specific quantum-AI applications in chemistry, optimization, and sensing are achieving practical utility in 2024-2025. This development marks a significant transition from theoretical potential to commercial realization, informed by rigorous evaluation methodologies.
Objective: To synthesize the state of the art on how quantum computing (QC) and artificial intelligence (AI) intersect, including algorithmic foundations, software/hardware stacks, empirical evidence of advantage, applications, and open challenges. Methods: We executed a structured search (2019 – 20th August 2025) across major journals (Nature portfolio, APS/PR, Elsevier), arXiv, and official framework documentation (Qiskit, Cirq/TFQ, PennyLane; AWS Braket; NVIDIA CUDA-Q). We included peer‑reviewed studies, comprehensive surveys, and official docs; we excluded opinion pieces and non-reproducible claims. Results: The field converges on hybrid quantum‑classical workflows with two dominant AI families: (i) quantum kernel methods and (ii) variational circuits/QNNs. Trainability remains a core challenge (barren plateaus and noise), with mitigation strategies advancing (local costs, layerwise training, engineered dissipation, learned decoders). Frameworks and cloud orchestrators now support end‑to‑end hybrid training. Empirical “utility” demonstrations in physics‑inspired tasks exist, while domain‑level advantage in mainstream ML is unproven and data‑dependent. Conclusions: QC‑AI integration is maturing into reproducible hybrid stacks. Near‑term value lies in quantum kernels with tailored feature maps, physics‑regularized QNNs, simulator‑accelerated pipelines, and AI‑for‑QC (calibration/decoding). Broad ML advantage awaits lower‑noise hardware and quantum‑friendly data embeddings.
Three recent experimental studies of D-Wave annealers each report an unexplained effective-temperature phenomenon: fitted inverse temperatures whose dependence on annealing time is left empirical, with sampling quality degrading at a critical frustration point [Pelofske, PRR 8, 033280 (2026)]; a coupling-independent temperature offset t-bar in [0.2,0.4] that violates the assumed T_eff ∝ 1/J scaling [Grattan et al., arXiv:2512.03162]; and hardware inverse temperatures a "consistent" factor of 5–7 above closed-system theory, attributed to uncontrollable hardware effects [Kim et al., PRE 113, 035302 (2026)]. We show that a single account, built from the diabatic effective-temperature relation of Gyhm et al. evaluated on the vendor's published fast-anneal schedules and one thermal bath, organizes all three. First, on real schedules the coherent inverse temperature is nearly linear in annealing time — not ∝ sqrt(t_a) as for linearly vanishing drivers — because the transverse field decays exponentially and dies by s ≈ 0.6; exact statevector simulation confirms the first-order integral to 0.4–4.2% in its validity window. Second, a units identity, beta_cyclic(t_a) = beta_correct(t_a/2pi), resolves the factor of 5–7: the published theoretical reference curves in this lane are quantitatively consistent with schedule data evaluated in cyclic rather than angular frequency units. The foundational paper's own schedule evaluations carry the rescale: its earlier single evaluation, beta ≈ 0.54 at 5 ns, equals our correct-convention value at t_a/2pi (0.539), and all 24 of its current tabulated beta and c_2 values that we can recompute from vendor schedule tables are reproduced at t_a/2pi to within 0.11%; a free-rescale fit to Kim et al.'s published simulation curve gives r = 6.40, with 2pi = 6.283 inside the fit's shallow region; what remains after the correction is consistent with open-system cooling. Third, the coupling-independent offset is a residual-excitation floor read through a domain-wall thermometer, t-bar ≈ 2/ln(1/rho): the implied kink densities (9e-5 to 4e-3) overlap published kink measurements, the excitation-floor form beats the constant-offset (noise) model on AIC for all four reported machines, and the logarithm compresses two and a half decades of rho into the narrow observed band, explaining its apparent universality. In the coherent nanosecond regime the closed-system law is an upper envelope on hardware inverse temperature; the shortfall is a candidate in-situ decoherence diagnostic, machine-resolved, and at the one matched cell where both machines are coherent the measured cross-machine ratio (2.63) agrees with the parameter-free prediction at the few-percent level. All hypotheses, tolerances, and kill criteria were registered before data extraction; failed and undecidable gates are reported alongside the successes. No quantum-advantage claim is made. AI use and validation: This work was produced in close collaboration with AI research assistants, directed by the author, who set the research direction, the standards of evidence, and all decisions about what is claimed and published. Anthropic's Claude (Fable 5) performed the analysis, simulation, extraction, and drafting; xAI's Grok (via Cursor) provided independent pre-publication review; and Anthropic's Claude Opus 5.5 performed a post-publication adversarial audit of the manuscript against its pinned evidence, with findings and dispositions logged in the project ledger and incorporated in subsequent versions. The registered-verdict protocol is designed so that no result rests on trusting a model's output: hypotheses, tolerances, and kill criteria were committed to the repository before any data was extracted; failed and undecidable gates are reported in the paper; every numeric claim traces to pinned, machine-readable extraction files and analysis code included in this deposit; figure data was extracted from the source PDFs' vector drawing operations rather than read by a model; and verbatim quotes were verified against raw page dumps.