Pith. sign in

REVIEW 3 major objections 6 minor 2 cited by

QAOA-GPT: Efficient Generation of Adaptive and Regular Quantum Approximate Optimization Algorithm Circuits

T0 review · 3 major / 6 minor · reviewed 2026-08-16 · deepseek-v4-flash

Pith's one-line read A GPT model trained on ADAPT-QAOA circuits can synthesize near-optimal MaxCut circuits in a single forward pass.

desk verdict Solid proof-of-concept for training a GPT to generate QAOA circuits, but the evaluation is not anchored to exact MaxCut and the missing artifacts weaken the reproducibility. read the letter →

arxiv 2504.16350 v1 pith:NOZ67NEQ submitted 2025-04-23 quant-ph cs.AI

classification quant-phcs.AI
keywords QAOAADAPT-QAOAgenerativepretrainedtransformerquantumcircuitgenerationMaxCutgraphembeddingsvariationaloptimizationCUDA-Q
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper introduces QAOA-GPT, a decoder-only transformer trained to output MaxCut quantum circuits directly from graph inputs. It aims to establish that a single-pass generative model can replace the iterative gradient-based loop of QAOA and ADAPT-QAOA without sacrificing solution quality: on unseen graphs of 10, 12, and 14 nodes, the best sampled circuit achieves mean approximation ratio 0.971-0.972, within 0.002-0.005 of ADAPT-QAOA's 0.973-0.974. Because inference is one forward pass, circuit generation runtime is nearly constant in problem size, in contrast to the steep growth of ADAPT-QAOA. This matters if the bottleneck of variational quantum optimization is the classical outer loop: a learned generator could make circuit synthesis cheap enough for routine use.

What carries the argument

The central object is the QAOA-GPT model itself: a decoder-only transformer (a GPT-2-style architecture trained from scratch) that predicts circuit tokens one at a time. Each training instance is a tokenized weighted edge list for a graph followed by tokenized layer blocks of an ADAPT-QAOA circuit, where each block records an operator index and the optimized parameters ($\gamma_k$, $\beta_k$). A FEATHER graph embedding (a characteristic-function-based structural descriptor) is broadcast into the input at every token position, letting the model condition circuit generation on global graph structure. The mechanism replaces the variational outer loop: at inference the model samples a full circuit in one forward pass, including both operator identities and parameter values, with no gradient evaluation or classical parameter optimization.

What would settle it

Compute exact MaxCut values (by brute force or an exact solver) for the same 1000 test graphs per size used in Table I and recompute the mean best approximation ratios of QAOA-GPT and ADAPT-QAOA; if the gap exceeds the reported 0.002-0.005 or QAOA-GPT's exact AR falls below 0.97, the central claim is refuted.

Watch

Extended reading notes

Core claim

The paper's discovery, on its own terms, is that a GPT-style transformer trained on ADAPT-QAOA-generated circuits can autoregressively synthesize problem-specific QAOA circuits for new MaxCut instances, jointly producing operator choices and variational parameters. The paper reports that the best of five sampled circuits for each graph tracks ADAPT-QAOA's approximation ratio to within a few thousandths (mean best AR 0.971-0.972 versus 0.973-0.974 across n = 10, 12, 14), while the generated circuits are comparable in layer count. It also reports that graph embeddings are necessary for out-of-distribution generalization, and that raising the target approximation ratio of the training circuits raises the model's output quality. The conclusion the authors draw is that generative inference can stand in for iterative variational optimization when near-optimal solutions are sufficient.

Load-bearing premise

The reported approximation ratios assume the classical heuristic used for OPT(G) returns the true MaxCut value; if it is inaccurate for the tested graphs, the quality labels on training circuits and the comparison with ADAPT-QAOA are both biased.

Editorial extensions

If this is right

  • For new MaxCut instances, QAOA-GPT generates circuits with no gradient evaluations or classical parameter optimization at inference, so circuit generation time remains nearly constant as problem size grows.
  • The target approximation ratio used to filter ADAPT-QAOA training circuits directly controls the model's output quality: models trained on circuits with higher target AR produce higher-AR circuits.
  • Incorporating graph embeddings improves performance on out-of-distribution graph families, with the largest gains on Barabasi-Albert and bipartite graphs.
  • Training on a mix of graph families (Erdos-Renyi and Barabasi-Albert) improves generalization to Barabasi-Albert graphs without hurting Erdos-Renyi performance.
  • When the target precision is tightened to $\alpha \ge 1 - 10^{-4}$ and a dual-qubit mixer pool is used, QAOA-GPT's approximation ratio plateaus near 0.8, showing that high-precision circuit generation is not yet solved.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The same graph-embedding-plus-edgelist conditioning could extend to other QUBO problems such as graph partitioning, community detection, or coloring, since the cost Hamiltonian enters only through the ADAPT-QAOA training labels.
  • If inference cost stays nearly constant while ADAPT-QAOA data-generation cost grows, the practical scaling bottleneck shifts to training-data generation; one testable prediction is that model AR should improve predictably with dataset size and diversity, following a scaling law.
  • Because the model emits both operators and parameters, its outputs could serve as a warm start for ADAPT-QAOA or conventional QAOA, potentially cutting the number of optimizer iterations needed for high precision.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 6 minor

Summary. The paper introduces QAOA-GPT, a decoder-only transformer trained to autoregressively generate quantum circuits for weighted MaxCut instances. The model conditions on tokenized graph edge lists plus FEATHER graph embeddings and outputs circuit operators and parameters in a single forward pass, bypassing the iterative gradient-based optimization used by ADAPT-QAOA. Training data are generated by ADAPT-QAOA with either the standard QAOA mixer pool or a dual-qubit Pauli pool, and circuits are filtered by an approximation-ratio threshold alpha >= 0.97. Experiments on 1000 unseen Erdős–Rényi graphs per size n=10, 12, 14 report mean best AR of 0.971–0.972, within roughly 0.002–0.005 of ADAPT-QAOA and above standard QAOA by about 0.03. The paper also includes ablations on graph embeddings, training distribution, target AR, operator pool choice, and runtime scalability.

Significance. If the numerical claims are supported, the paper is a useful proof-of-concept that generative transformers can synthesize QAOA-like circuits with quality close to adaptive variational methods while avoiding iterative optimization at inference time. The evaluation against an external approximation-ratio objective, rather than only against training loss, mitigates the most obvious form of supervised-imitation circularity, and the FEATHER-conditioning ablation is informative. The GPU-accelerated data-generation scaling results are also relevant to practical use. The quantitative significance, however, is not yet established: the reported approximation ratios depend on a heuristic MaxCut value that is never checked against exact optima for the small tested sizes, and the best-of-five reporting without per-graph uncertainty makes the headline margins under-specified. No code or data are released, which further limits reproducibility.

major comments (3)
  1. [Section III, Step 1; Table I; Figure 3] The approximation ratio is defined as alpha = <H_c>/OPT(G), where OPT(G) is described as 'the MaxCut value obtained via classical heuristics [25]' (Section III, Step 1). This heuristic value is used twice: it filters the ADAPT-QAOA training circuits at alpha >= 0.97, and it is the denominator of every AR reported in Table I and Figure 3. For the evaluated sizes n=10, 12, and 14, exact MaxCut can be computed by enumerating all 2^n cuts, yet the paper does not verify the heuristic against exact values. If reference [25] returns feasible cuts (a lower bound), all reported ARs are inflated and the gap to ADAPT-QAOA is overstated; if it returns a relaxation upper bound, the ARs are depressed. The headline magnitudes '0.971 to 0.972' and 'within 0.002 to 0.005' are therefore not anchored to ground truth. Please recompute every reported AR using exact OPT for these sizes, and re-examine the training-data filter in light of the exact values.
  2. [Section IV-B; Table I; Figure 3] The paper reports QAOA-GPT 'best AR' as the maximum over five generated circuits per graph, but the figures and table present only the mean over graphs with a standard deviation. No confidence intervals on the mean, no distribution over the five generated circuits, and no paired statistical comparison with ADAPT-QAOA are given. Because the claimed margin to ADAPT-QAOA is only 0.002–0.005 in AR, selection noise from best-of-five reporting can affect the stated ranking, and the standard deviation across graphs in Table I does not quantify this. Please report the per-graph best-of-five distribution (mean, worst-of-five, spread across seeds) and a confidence interval or paired test for the ADAPT-QAOA and standard-QAOA comparisons.
  3. [Section V-F; Figure 7] The runtime comparison in Figure 7 compares a single forward pass of the trained QAOA-GPT with the entire iterative ADAPT-QAOA optimization, which is appropriate for the paper's speed claim, but the measurement protocol is underspecified: hardware, batch size, number of ADAPT-QAOA layers or optimizer iterations, and any equal-quality criterion are not stated. The text asserts 'exponential scaling' for ADAPT-QAOA, but the figure shows no raw data, error bars, or fitted scaling. Please specify the protocol, report repeated measurements, and clarify what circuit quality is held fixed when comparing inference time with iterative optimization time.
minor comments (6)
  1. [Abstract] The sentence 'demonstrate that QAOA-GPT, generates high quality quantum circuits' contains an extra comma after 'QAOA-GPT'.
  2. [Section V] The opening phrase 'In the this section' has a duplicated article and should read 'In this section'.
  3. [Table I] The table header is ambiguous: the columns labeled 'QAOA-GPT (Best)' and 'QAOA-GPT' under both AR and number of layers do not state which is average and which is best-of-five; Figure 3 distinguishes 'Avg. AR' and 'Best AR', so the table should use matching labels.
  4. [Section IV-B] The text says 'Each data point represents the mean AR over 50 random graphs within the corresponding density bin' and 'A total of 1000 random graphs were evaluated for each problem size.' If densities range from 0.3 to 0.9 in increments of 0.1, seven bins times fifty graphs gives 350 graphs, not 1000; please clarify the binning scheme.
  5. [Section V-G] There are several typos in this section: 'noisless' should be 'noiseless', 'statevctor' should be 'statevector', and 'scinario' should be 'scenario'.
  6. [Figure 6] The caption 'target = 1 10 4' should be typeset as a mathematical expression, e.g., alpha >= 1 - 10^{-4}, to be readable.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: QAOA-GPT is supervised imitation evaluated on held-out graphs with an external approximation-ratio objective; the one self-citation affecting setup is not load-bearing.

full rationale

QAOA-GPT is an empirical supervised-learning pipeline, not a derivation. The training circuits are produced by ADAPT-QAOA and filtered by the target ratio α = ⟨ψ|Hc|ψ⟩/OPT(G), where OPT(G) is 'the MaxCut value obtained via classical heuristics [25]'; the transformer is trained with a next-token prediction loss only. The reported results (Table I: best AR 0.971–0.972 vs ADAPT-QAOA 0.973–0.974 on n=10,12,14) are obtained by simulating the generated circuits and dividing the expectation value by the same per-graph OPT(G), so the objective is an external benchmark rather than an input to the model. The model imitates its teacher, which explains why its AR is near the 0.97 training threshold, but matching a teacher on held-out inputs is the standard goal of imitation learning, not a circular reduction. The paper does contain self-citations, notably [26] ('the required QAOA depth rarely exceeds the number of nodes... therefore we set the depth of QAOA circuits equal to problem size p=n'), which is used only to fix the standard-QAOA baseline depth, and [27,28] as motivation for graph embeddings; neither enters the QAOA-GPT accuracy computation, and the embedding benefit is tested by an in-paper ablation (Table II). A robustness caveat, which is a correctness issue rather than circularity, is that ARs are normalized by a heuristic OPT rather than exact MaxCut values; this could bias the reported magnitudes but does not make the claim self-referential.

Assumptions & free parameters 5 free parameters · 6 assumptions · 0 invented entities

The paper is an empirical ML system with no derivation. The main free parameters are the training AR threshold, gamma0 grid, QAOA depth, number of samples, and undisclosed model hyperparameters. The axioms are standard reductions plus domain assumptions that the teacher signal, the classical OPT estimate, the graph embedding, and the surrogate loss are adequate.

free parameters (5)
  • Training AR filter threshold = 0.97
    Circuits with approximation ratio below 0.97 are excluded from the training set; Section V-C shows model AR scales with this threshold.
  • ADAPT-QAOA initial gamma0 grid = {0.01, 0.1, 0.5, 1}
    Grid searched to enrich training data; no principled selection is given, and it affects circuit diversity.
  • Standard QAOA depth p = p = n
    Depth set equal to number of nodes following [26]; this affects the QAOA baseline AR.
  • Number of generated circuits per test graph = 5
    Best-of-5 and average AR are reported; best AR is inflated relative to a single sample.
  • QAOA-GPT hyperparameters
    Model size, layers, heads, learning rate, context window T, bmin/bmax, and FEATHER embedding dimension are not reported; these hand-chosen settings directly determine the results and are needed for reproduction.
assumptions (6)
  • standard math MaxCut can be represented as a QUBO and mapped to cost Hamiltonian H_c = -1/2 sum w_ij (I - Z_i Z_j).
    Standard reduction in Section II-A, used for all AR calculations.
  • domain assumption ADAPT-QAOA gradient-based operator selection produces high-quality circuits.
    Used to generate training labels; if ADAPT-QAOA circuits are not near-optimal or not diverse, the teacher signal is poor (Sections III and IV-A).
  • domain assumption The classical heuristic [25] gives accurate OPT(G) values.
    OPT(G) defines the approximation ratio threshold and all reported ARs; no exact MaxCut verification for n up to 14 is provided.
  • domain assumption FEATHER embeddings plus edge-list tokens contain sufficient information to predict circuit structure.
    The model conditions on FEATHER embeddings; if they miss relevant graph structure, generalization fails. The ablation in Table II supports this empirically but does not prove sufficiency.
  • domain assumption Autoregressive next-token prediction on tokenized circuits is an appropriate surrogate for circuit quality.
    The training loss does not include AR; the model is only trained to imitate high-quality circuits, so it may not learn to avoid low-quality circuits (Section V-D).
  • domain assumption Statevector simulation gives the true expectation value for generated circuits.
    All ARs are computed in noiseless simulation; hardware noise and sampling would degrade performance.

how reviews work

0 comments
Cite this review

Pith. "Pith review of QAOA-GPT: Efficient Generation of Adaptive and Regular Quantum Approximate Optimization Algorithm Circuits." pith.science (2026). https://pith.science/paper/NOZ67NEQ

@misc{pith2026250416350,
  author       = {Pith},
  title        = {Pith review of: QAOA-GPT: Efficient Generation of Adaptive and Regular Quantum Approximate Optimization Algorithm Circuits},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/NOZ67NEQ}},
  note         = {Machine review of arXiv:2504.16350}
}
read the original abstract

Quantum computing has the potential to improve our ability to solve certain optimization problems that are computationally difficult for classical computers, by offering new algorithmic approaches that may provide speedups under specific conditions. In this work, we introduce QAOA-GPT, a generative framework that leverages Generative Pretrained Transformers (GPT) to directly synthesize quantum circuits for solving quadratic unconstrained binary optimization problems, and demonstrate it on the MaxCut problem on graphs. To diversify the training circuits and ensure their quality, we have generated a synthetic dataset using the adaptive QAOA approach, a method that incrementally builds and optimizes problem-specific circuits. The experiments conducted on a curated set of graph instances demonstrate that QAOA-GPT, generates high quality quantum circuits for new problem instances unseen in the training as well as successfully parametrizes QAOA. Our results show that using QAOA-GPT to generate quantum circuits will significantly decrease both the computational overhead of classical QAOA and adaptive approaches that often use gradient evaluation to generate the circuit and the classical optimization of the circuit parameters. Our work shows that generative AI could be a promising avenue to generate compact quantum circuits in a scalable way.

Figures

Figures reproduced from arXiv: 2504.16350 by the authors.

Figure 1
Figure 1. Schematic representation of the QAOA-GPT framework. Random [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. Proposed use case diagram. Given a user-supplied input graph, the [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Average approximation ratio (AR) of quantum circuits generated [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: QAOA-GPT performance as a function of the target approximation [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]
Figure 5
Figure 5. Figure 5: QAOA-GPT performance over the course of training . The blue curve [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]
Figure 7
Figure 7. Figure 7: Running time scalability with problem size. The plot shows that [PITH_FULL_IMAGE:figures/full_fig_p009_7.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Learning to Prepare Molecular Ground States with Transformer Models

    quant-ph 2026-07 conditional novelty 6.0 of 10

    Transformers trained on ADAPT-VQE data generate imipramine ground-state circuits in seconds at roughly reference accuracy — and beat the training data after reinforcement learning — though real-hardware energies still...

  2. Performance Model for Hybrid Quantum-Classical Workflows

    quant-ph 2026-07 conditional novelty 6.0 of 10

    A two-level runtime model decomposes hybrid quantum-classical cycles into quantum, classical, and communication time, allowing a communication-to-computation ratio to classify workflows as compute- or communication-bound.

Reference graph

Works this paper leans on

35 extracted references · 23 canonical work pages · cited by 2 Pith papers

  1. [25]

    Rank-two relaxation heuristics for max-cut and other binary quadratic programs

    Samuel Burer, Renato Monteiro, and Yin Zhang. Rank-two relaxation heuristics for max-cut and other binary quadratic programs. SIAM Journal on Optimization , 12, 07 2001

  2. [1]

    Quantum computing for finance

    Dylan Herman, Cody Googin, Xiaoyuan Liu, Yue Sun, Alexey Galda, Ilya Safro, Marco Pistoia, and Yuri Alexeev. Quantum computing for finance. Nature Reviews Physics , 5(8):450–465, 2023

  3. [2]

    Quantum chemistry in the age of quantum computing

    Yudong Cao, Jonathan Romero, Jonathan P Olson, Matthias Degroote, Peter D Johnson, M ´aria Kieferov´a, Ian D Kivlichan, Tim Menke, Borja Peropadre, Nicolas PD Sawaya, Sukin Sim, Libor Veis, and Alan Aspuru-Guzik. Quantum chemistry in the age of quantum computing. Chemical reviews, 119(19):10856–10915, 2019

  4. [3]

    Quantum-centric supercomputing for materials science: A perspective on challenges and future directions

    Yuri Alexeev, Maximilian Amsler, Marco Antonio Barroca, Sanzio Bassini, Torey Battelle, Daan Camps, and et al. Quantum-centric supercomputing for materials science: A perspective on challenges and future directions. Future Generation Computer Systems , 160:666–710, 2024

  5. [4]

    Ruslan Shaydulin, Hayato Ushijima-Mwesigwa, Christian F. A. Negre, Ilya Safro, Susan M. Mniszewski, and Yuri Alexeev. A hybrid approach for solving optimization problems on small quantum computers. Com- puter, 52(6):18–26, June 2019

  6. [5]

    Quantum machine learning

    Jacob Biamonte, Peter Wittek, Nicola Pancotti, Patrick Rebentrost, Nathan Wiebe, and Seth Lloyd. Quantum machine learning. Nature, 549(7671):195–202, 2017

  7. [6]

    A quantum approximate optimization algorithm

    Edward Farhi, Jeffrey Goldstone, and Sam Gutmann. A quantum approximate optimization algorithm. arXiv preprint arXiv:1411.4028 , 2014

  8. [7]

    A review on quantum approximate optimization algorithm and its variants

    Kostas Blekos, Dean Brand, Andrea Ceschini, Chiao-Hui Chou, Rui- Hao Li, Komal Pandya, and Alessandro Summer. A review on quantum approximate optimization algorithm and its variants. Physics Reports , 1068:1–66, 2024

Show all 35 references
  1. [8]

    Adaptive quantum approximate optimization algorithm for solving combinato- rial problems on a quantum computer

    Linghua Zhu, Ho Lun Tang, George S Barron, FA Calderon-Vargas, Nicholas J Mayhall, Edwin Barnes, and Sophia E Economou. Adaptive quantum approximate optimization algorithm for solving combinato- rial problems on a quantum computer. Physical Review Research , 4(3):033029, 2022

  2. [9]

    An adaptive variational algorithm for exact molecular simula- tions on a quantum computer

    Harper R Grimsley, Sophia E Economou, Edwin Barnes, and Nicholas J Mayhall. An adaptive variational algorithm for exact molecular simula- tions on a quantum computer. Nature communications, 10(1):1–9, 2019

  3. [10]

    Layer VQE: A variational approach for combinatorial optimization on noisy quantum computers

    Xiaoyuan Liu, Anthony Angone, Ruslan Shaydulin, Ilya Safro, Yuri Alexeev, and Lukasz Cincio. Layer VQE: A variational approach for combinatorial optimization on noisy quantum computers. IEEE Transactions on Quantum Engineering , 3:1–20, 2022

  4. [11]

    Grovergpt: A large language model with 8 billion parameters for quantum searching

    Haoran Wang, Pingzhi Li, Min Chen, Jinglei Cheng, Junyu Liu, and Tianlong Chen. Grovergpt: A large language model with 8 billion parameters for quantum searching. arXiv preprint arXiv:2501.00135 , 2024

  5. [12]

    The generative quantum eigensolver (gqe) and its application for ground state search

    Kouhei Nakaji, Lasse Bjørn Kristensen, Jorge A Campos-Gonzalez- Angulo, Mohammad Ghazi Vakili, Haozhe Huang, Mohsen Bagher- imehrab, Christoph Gorgulla, FuTe Wong, Alex McCaskey, Jin-Sung Kim, et al. The generative quantum eigensolver (gqe) and its application for ground state...

  6. [13]

    cuquantum sdk: A high-performance library for accelerating quantum science

    Harun Bayraktar, Ali Charara, David Clark, Saul Cohen, Timothy Costa, Yao-Lung L Fang, Yang Gao, Jack Guan, John Gunnels, Azzam Haidar, et al. cuquantum sdk: A high-performance library for accelerating quantum science. In 2023 IEEE International Conference on Quantum Computing...

  7. [14]

    NVIDIA CUDA-Q framework

    NVIDIA Corporation. NVIDIA CUDA-Q framework. https://github. com/NVIDIA/cuda-quantum

  8. [15]

    Some simplified np-complete problems

    Michael R Garey, David S Johnson, and Larry Stockmeyer. Some simplified np-complete problems. In Proceedings of the sixth annual ACM symposium on Theory of computing , pages 47–63, 1974

  9. [16]

    Quantum bridge analytics i: a tutorial on formulating and using qubo models

    Fred Glover, Gary Kochenberger, Rick Hennig, and Yu Du. Quantum bridge analytics i: a tutorial on formulating and using qubo models. Annals of Operations Research , 314(1):141–183, 2022

  10. [17]

    A tutorial on formulating qubo models

    Fred Glover and Gary Kochenberger. A tutorial on formulating qubo models. arXiv preprint arXiv:1811.11538 , 2018

  11. [18]

    Multilevel combi- natorial optimization across quantum architectures

    Hayato Ushijima-Mwesigwa, Ruslan Shaydulin, Christian FA Negre, Susan M Mniszewski, Yuri Alexeev, and Ilya Safro. Multilevel combi- natorial optimization across quantum architectures. ACM Transactions on Quantum Computing , 2(1):1–29, 2021

  12. [19]

    Hybrid quantum-classical algorithms for approximate graph coloring

    Sergey Bravyi, Alexander Kliesch, Robert Koenig, and Eugene Tang. Hybrid quantum-classical algorithms for approximate graph coloring. Quantum, 6:678, 2022

  13. [20]

    Equivariant qaoa and the duel of the mixers

    Boris Tsvelikhovskiy, Ilya Safro, and Yuri Alexeev. Equivariant qaoa and the duel of the mixers. arXiv preprint arXiv:2405.07211 , 2024

  14. [21]

    Improving language understanding by generative pre-training

    Alec Radford, Karthik Narasimhan, Tim Salimans, Ilya Sutskever, et al. Improving language understanding by generative pre-training. 2018

  15. [22]

    Attention is all you need

    Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, Łukasz Kaiser, and Illia Polosukhin. Attention is all you need. In Advances in neural information processing systems , pages 5998–6008, 2017

  16. [23]

    Evaluating large language models trained on code

    Mark Chen, Jerry Tworek, Heewoo Jun, Qiming Yuan, Henrique Ponde De Oliveira Pinto, Jared Kaplan, Harri Edwards, Yuri Burda, Nicholas Joseph, Greg Brockman, et al. Evaluating large language models trained on code. arXiv preprint arXiv:2107.03374 , 2021

  17. [24]

    Characteristic functions on graphs: Birds of a feather, from statistical descriptors to parametric models

    Benedek Rozemberczki and Rik Sarkar. Characteristic functions on graphs: Birds of a feather, from statistical descriptors to parametric models. In Proceedings of the 29th ACM international conference on information & knowledge management , pages 1325–1334, 2020

  18. [26]

    Classi- cal symmetries and the quantum approximate optimization algorithm

    Ruslan Shaydulin, Stuart Hadfield, Tad Hogg, and Ilya Safro. Classi- cal symmetries and the quantum approximate optimization algorithm. Quantum Information Processing , 20(11):1–28, 2021

  19. [27]

    Graph repre- sentation learning for parameter transferability in quantum approximate optimization algorithm

    Jose Falla, Quinn Langfitt, Yuri Alexeev, and Ilya Safro. Graph repre- sentation learning for parameter transferability in quantum approximate optimization algorithm. Quantum Machine Intelligence , 6(2):46, 2024

  20. [28]

    Similarity-based parameter transferability in the quantum approximate optimization algorithm

    Alexey Galda, Eesh Gupta, Jose Falla, Xiaoyuan Liu, Danylo Lykov, Yuri Alexeev, and Ilya Safro. Similarity-based parameter transferability in the quantum approximate optimization algorithm. Frontiers in Quantum Science and Technology , 2, 2023

  21. [29]

    Hagberg, D.A

    A. Hagberg, D.A. Schult, and P.J. Swart. Exploring network structure, dynamics, and function using networkx. In Proceedings of the 7th Python in Science Conference SciPy2008 , volume 836, pages 11–15, 2008

  22. [30]

    Generative quantum combinatorial optimization by means of a novel conditional generative quantum eigensolver

    Shunya Minami, Kouhei Nakaji, Yohichi Suzuki, Al ´an Aspuru-Guzik, and Tadashi Kadowaki. Generative quantum combinatorial optimization by means of a novel conditional generative quantum eigensolver. arXiv preprint arXiv:2501.16986, 2025

  23. [31]

    NVIDIA CUDA-Q

    NVIDIA Corporation. NVIDIA CUDA-Q. https://developer.nvidia.com/ cuda-q, 2025. Accessed: 2025-04-09

  24. [32]

    NVIDIA DGX Quantum

    NVIDIA Corporation. NVIDIA DGX Quantum. https://www.nvidia. com/en-us/data-center/dgx-quantum/, 2025. Accessed: 2025-04-09

  25. [33]

    Mlqaoa: Graph learning ac- celerated hybrid quantum-classical multilevel qaoa

    Bao Bach, Jose Falla, and Ilya Safro. Mlqaoa: Graph learning ac- celerated hybrid quantum-classical multilevel qaoa. In 2024 IEEE International Conference on Quantum Computing and Engineering (QCE), volume 01, pages 1–12, 2024

  26. [34]

    Scal- ing up the quantum divide and conquer algorithm for combinatorial optimization

    Cameron Ibrahim, Teague Tomesh, Zain Saleem, and Ilya Safro. Scal- ing up the quantum divide and conquer algorithm for combinatorial optimization. In 2024 IEEE International Conference on Quantum Computing and Engineering (QCE) , volume 1, pages 545–551. IEEE,

  27. [2024]

    arXiv preprint arXiv:2405.00861

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.