Pith. sign in

REVIEW 3 major objections 1 minor 1 cited by

The loss tolerance of cat breeding for fault-tolerant grid state generation

T0 review · 3 major / 1 minor · reviewed 2026-08-05 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read Cat breeding cannot produce fault-tolerant GKP states once total optical loss exceeds 4%.

desk verdict The supplied full text is an unrelated computer-vision dataset paper, so the abstract's 4% loss threshold for cat breeding is completely unsupported in this artifact. read the letter →

arxiv 2508.06193 v1 pith:A62S6PIF submitted 2025-08-08 quant-ph

Olga Solodovnikova , Ulrik L. Andersen , Jonas S. Neergaard-Nielsen This is my paper
classification quant-ph
keywords catbreedingGKPstatescontinuous-variablequantumcomputingopticallossWignerfunctionGaussian-mixturesimulationfault-tolerancethresholdhomodynedetection
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 targets the cat breeding protocol, currently the leading way to deterministically prepare Gottesman-Kitaev-Preskill (GKP) states for continuous-variable photonic quantum computers. It claims that optical loss lowers the protocol's success probability and, above a total loss of 4%, prevents preparation of a GKP state of fault-tolerant quality. The load-bearing step is a simulation that writes each input's Wigner function as a linear combination of Gaussians, making multi-round breeding with lossy, mixed states computationally tractable. If the claim holds, experimental builders get a concrete loss budget: keep total loss below about 4% or the prepared GKP states will not be good enough for fault-tolerant operation.

What carries the argument

The Gaussian-mixture Wigner representation: each input state's Wigner function is written as a linear combination of Gaussian terms, and the protocol's linear-optics operations—beam splitters, homodyne detection, feedforward displacement—together with loss act on this mixture round by round. This avoids the exponential scaling that ordinarily makes multi-round breeding with lossy inputs hard to analyze, and it is what lets the simulation produce the 4% loss threshold.

What would settle it

Run the published simulator at total losses of 2%, 3%, 4%, and 5% and compare the GKP quality against an independent calculation or a calibrated experiment that does not rely on the Gaussian-mixture truncation; if quality still crosses the fault-tolerance threshold above 4%, the claimed budget is wrong.

Watch

Extended reading notes

Core claim

On the paper's own terms: representing the Wigner function of squeezed cat states as a sum of Gaussians allows the cat breeding protocol—interfering cat states on beam splitters, homodyne detecting, and feeding forward a displacement—to be simulated through several rounds even when the inputs are mixed by loss. Running this simulation shows that optical loss reduces the overall success probability of the protocol and that when total loss exceeds 4% the resulting GKP state no longer meets the quality required for fault tolerance. The method is released as open-source code, so the threshold can be reproduced and explored.

Load-bearing premise

The 4% loss threshold stands or falls with the accuracy of the Gaussian-mixture approximation after many breeding rounds, especially how truncation error behaves as loss increases, and with the fault-tolerance criterion chosen for the GKP state.

Editorial extensions

If this is right

  • Experiments using cat breeding must keep total optical loss under roughly 4% to produce fault-tolerant GKP states.
  • The 4% budget applies to the whole chain—cat source, beam splitters, detectors, feedforward—so component losses have to be engineered against a shared target.
  • The Gaussian-mixture simulation gives a practical way to choose the number of breeding rounds and beam-splitter ratios under realistic loss.
  • The open-source code lets other groups reproduce the threshold and benchmark their own GKP preparation approaches against it.

Reading between the lines

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

  • The abstract leaves the truncation details of the Gaussian-mixture expansion unspecified; if truncation error grows with loss, the exact position of the 4% threshold could move, so an independent high-loss verification would be valuable.
  • The meaning of 'fault-tolerant GKP' depends on the chosen quality metric, so the 4% number is tied to that criterion; a stricter or looser benchmark would shift the loss budget.
  • The same Gaussian-mixture machinery likely transfers to other continuous-variable state-preparation protocols, making the simulation method potentially broader than the specific cat-breeding result.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 1 minor

Summary. The arXiv listing (2508.06193) claims a quantitative study of the cat-breeding protocol for GKP state preparation under optical loss. The abstract states that by representing Wigner functions as linear combinations of Gaussians, several rounds of breeding with mixed input states can be simulated quickly and accurately, and that optical loss prohibits preparation of a fault-tolerant GKP state when loss exceeds 4%. The abstract also promises open-source code. However, the supplied full text is arXiv:2508.06205, "PA-HOI: A Physics-Aware Human and Object Interaction Dataset," a computer-vision paper about human-object interaction motion capture. The body contains no mention of GKP states, cat breeding, beam splitters, homodyne detection, optical loss, or any quantum simulation. Consequently, the reviewed artifact consists of an abstract whose central claim is completely unsupported by the accompanying text.

Significance. If the abstract's result were backed by a rigorous derivation and reproducible simulation, the claimed 4% total-loss threshold for fault-tolerant GKP preparation would be a useful quantitative design guide for continuous-variable photonic quantum computing. The Gaussian-mixture Wigner-function method, if it indeed controls truncation error across multiple breeding rounds, would also be a methodological contribution. The promise of open-source code is commendable. However, none of these elements appear in the supplied full text. There are no derivations, no simulation details, no figures with numerical results, and no code artifact. The significance of the paper therefore cannot be assessed from the material under review; the only evidence of the claimed contribution is the abstract itself.

major comments (3)
  1. [Full text (entire body)] The supplied full text is a different paper entirely: 'PA-HOI: A Physics-Aware Human and Object Interaction Dataset' (arXiv:2508.06205), not the claimed quant-ph manuscript on cat breeding. The body contains no derivation of the central claim, no simulation description, no loss model, no GKP quality metric, and no numerical data supporting the 4% threshold. Under the review rule that all manuscript text is in-scope evidence, this mismatch is load-bearing: the abstract's quantitative result is entirely unsupported by the reviewed artifact.
  2. [Abstract] Even if the full text mismatch is set aside, the abstract alone is insufficient to validate the central claim. The Gaussian-mixture simulation is described only qualitatively: no truncation cutoff, no number of retained Gaussians per round, no error bound, and no specification of where loss is applied (input states, beam splitters, detection, or all). The 4% threshold could move with any of these choices, so the central quantitative result cannot be checked or reproduced from the claimed methodology as stated.
  3. [Abstract] The fault-tolerance criterion is not defined. The abstract says loss 'prohibits the preparation of a fault-tolerant GKP state when the loss exceeds 4%,' but does not state the quality threshold (e.g., effective squeezing in dB, or a specific error-correction threshold) used to classify a state as fault-tolerant. Without an externally fixed benchmark, there is a risk that the threshold is calibrated to the simulation's own success definition, which would make the conclusion circular. This concern cannot be resolved from the supplied text.
minor comments (1)
  1. [PA-HOI full text, §1] The unrelated full text contains typos such as 'scenarions' and incomplete reference formatting. These are not material to the quantum claim but further indicate that the body does not correspond to the submitted abstract.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity demonstrated: the supplied full text is a different paper, so the claimed derivation chain is absent from the artifact.

full rationale

The abstract of arXiv:2508.06193 claims that representing input Wigner functions as linear combinations of Gaussians makes it possible to 'quickly and accurately simulate several rounds of breeding' and that optical loss 'prohibits the preparation of a fault-tolerant GKP state when the loss exceeds 4%.' However, the supplied full text is arXiv:2508.06205, 'PA-HOI: A Physics-Aware Human and Object Interaction Dataset,' a computer-vision paper containing no mention of GKP states, cat breeding, Wigner functions, beam splitters, homodyne detection, or optical loss. There is therefore no derivation chain in the artifact to walk, and no equation or fitted parameter can be quoted that reduces the claimed 4% threshold to an input by construction. The absence of simulation details, truncation-error analysis, and an externally fixed fault-tolerance criterion is a completeness/correctness concern, not a circularity concern. No self-citation, ansatz smuggling, or renaming of a known result is present in the supplied body. Under the hard rule that circularity may be claimed only when the paper's own text exhibits a specific reduction, the honest verdict is no demonstrated circularity, score 0.

Assumptions & free parameters 2 free parameters · 3 assumptions · 0 invented entities

Because the submitted body text is an unrelated manuscript, the ledger is reconstructed from the abstract alone. No invented entities appear. The load-bearing assumptions are the Gaussian-mixture representation remaining valid under loss, the loss model's placement, and the external fault-tolerance criterion. Two implicit choices (the truncation cutoff for the mixture and the GKP quality criterion) function as free parameters in the sense that different values would change the reported 4%, but neither value is stated anywhere in the available text.

free parameters (2)
  • Gaussian-mixture truncation cutoff (terms kept per breeding round)
    The abstract says the Wigner function is represented as a linear combination of Gaussians to avoid exponential scaling; keeping all terms would still scale exponentially, so a cutoff is implied. The cutoff value, chosen by hand, sets the accuracy of the 4% threshold but is not stated in the abstract.
  • Fault-tolerance quality threshold (e.g., required GKP effective squeezing in dB)
    The 'fault-tolerant' verdict requires a metric or criterion against which the simulated GKP state is judged. The abstract names no value; the 4% loss number is meaningless without it.
assumptions (3)
  • domain assumption Photon loss acts as a Gaussian operation, so a sum-of-Gaussians Wigner representation remains a sum of Gaussians throughout the breeding protocol
    This is what makes the claimed simulation tractable; it is not stated in the abstract and cannot be checked because the body is missing.
  • domain assumption The fault-tolerance criterion for GKP states is a fixed, externally justified benchmark
    The 4% threshold is defined relative to this criterion; if the criterion were chosen to match the simulation, the result would be circular.
  • domain assumption Homodyne detection and feedforward displacement in the breeding protocol are modeled with stated (or ideal) efficiencies
    The abstract reports a loss threshold, so the loss model must specify where loss enters (input state, beam splitters, detection); the abstract does not say.

how reviews work

0 comments
Cite this review

Pith. "Pith review of The loss tolerance of cat breeding for fault-tolerant grid state generation." pith.science (2026). https://pith.science/paper/A62S6PIF

@misc{pith2026250806193,
  author       = {Pith},
  title        = {Pith review of: The loss tolerance of cat breeding for fault-tolerant grid state generation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/A62S6PIF}},
  note         = {Machine review of arXiv:2508.06193}
}
read the original abstract

The development of a continuous-variable photonic quantum computer depends on the reliable preparation of high-quality Gottesman-Kitaev-Preskill states. The most promising GKP preparation scheme is the cat breeding protocol, which can generate GKP states deterministically given a source of squeezed cat states, using beam splitters, homodyne detectors and a feedforward displacement. However, analyzing the performance of the protocol under loss is cumbersome due to the exponential scaling of the system. By representing the Wigner function of the input states as a linear combination of Gaussians, we are able to quickly and accurately simulate several rounds of breeding with mixed input states. Using this novel method, we find that optical loss decreases the overall success probability of the protocol, and prohibits the preparation of a fault-tolerant GKP state when the loss exceeds 4\%. Our methodology is available as open-source code.

Discussion (0). Sign in to comment.

Forward citations

Cited by 1 Pith paper

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

  1. Iterative $C_Z$-gate-based protocol for squeezed Schr\"odinger cat state engineering

    quant-ph 2026-06 unverdicted novelty 6.0 of 10

    A new iterative CZ-gate and homodyne protocol generates high-fidelity squeezed Schrödinger cat states with controllable size, squeezing, and tunable fidelity-success trade-off.

Reference graph

Works this paper leans on

40 extracted references · 30 canonical work pages · cited by 1 Pith paper

  1. [1]

    Bharat Lal Bhatnagar, Xianghui Xie, Ilya A Petrov, Cristian Sminchisescu, Chris- tian Theobalt, and Gerard Pons-Moll. 2022. BEHAVE: Dataset and method for tracking human object interactions. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition . 15935–15946

    work page 2022

  2. [2]

    Junuk Cha, Jihyeon Kim, Jae Shin Yoon, and Seungryul Baek. 2024. Text2hoi: Text-guided 3d motion generation for hand-object interaction. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition . 1577–1585

    work page 2024

  3. [3]

    Wenshuo Chen, Haozhe Jia, Songning Lai, Keming Wu, Hongru Xiao, Lijie Hu, and Yutao Yue. 2025. Free-T2M: Frequency Enhanced Text-to-Motion Diffusion Model With Consistency Loss. arXiv preprint arXiv:2501.18232 (2025)

    arXiv 2025

  4. [4]

    Xin Chen, Biao Jiang, Wen Liu, Zilong Huang, Bin Fu, Tao Chen, and Gang Yu. 2023. Executing your Commands via Motion Diffusion in Latent Space. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 18000–18010

    work page 2023

  5. [5]

    Peishan Cong, Ziyi Wang, Yuexin Ma, and Xiangyu Yue. 2025. SemGeoMo: Dynamic Contextual Human Motion Generation with Semantic and Geometric Guidance. arXiv preprint arXiv:2503.01291 (2025)

    work page Pith review arXiv 2025

  6. [6]

    Zicong Fan, Omid Taheri, Dimitrios Tzionas, Muhammed Kocabas, Manuel Kauf- mann, Michael J Black, and Otmar Hilliges. 2023. ARCTIC: A dataset for dexterous bimanual hand-object manipulation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition . 12943–12954

    work page 2023

  7. [7]

    Chuan Guo, Shihao Zou, Xinxin Zuo, Sen Wang, Wei Ji, Xingyu Li, and Li Cheng

  8. [8]

    Mohamed Hassan, Duygu Ceylan, Ruben Villegas, Jun Saito, Jimei Yang, Yi Zhou, and Michael J Black. 2021. Stochastic scene-aware motion prediction. In Proceedings of the IEEE/CVF International Conference on Computer Vision . 11374– 11384

    work page 2021

Show all 40 references
  1. [9]

    Seokhyeon Hong, Chaelin Kim, Serin Yoon, Junghyun Nam, Sihun Cha, and Junyong Noh. 2025. SALAD: Skeleton-aware Latent Diffusion for Text-driven Motion Generation and Editing. arXiv preprint arXiv:2503.13836 (2025)

    2025 arXiv

  2. [10]

    Liangxiao Hu, Hongwen Zhang, Yuxiang Zhang, Boyao Zhou, Boning Liu, Sheng- ping Zhang, and Liqiang Nie. 2024. Gaussianavatar: Towards realistic human avatar modeling from a single video via animatable 3d gaussians. In Proceedings of the IEEE/CVF conference on computer vision a...

    2024

  3. [11]

    Yinghao Huang, Omid Taheri, Michael J Black, and Dimitrios Tzionas. 2022. InterCap: Joint markerless 3D tracking of humans and objects in interaction. In DAGM German Conference on Pattern Recognition . Springer, 281–299

    2022

  4. [12]

    Yiheng Huang, Hui Yang, Chuanchen Luo, Yuxi Wang, Shibiao Xu, Zhaoxiang Zhang, Man Zhang, and Junran Peng. 2024. Stablemofusion: Towards robust and efficient diffusion-based motion generation framework. In Proceedings of the 32nd ACM International Conference on Multimedia . 224–232

    2024

  5. [13]

    Biao Jiang, Xin Chen, Wen Liu, Jingyi Yu, Gang Yu, and Tao Chen. 2023. Mo- tiongpt: Human motion as a foreign language. Advances in Neural Information Processing Systems 36 (2023), 20067–20079

    2023

  6. [14]

    Hyeonwoo Kim, Sookwan Han, Patrick Kwon, and Hanbyul Joo. 2025. Beyond the Contact: Discovering Comprehensive Affordance for 3D Objects from Pre- trained 2D Diffusion Models. In Computer Vision – ECCV 2024 , Aleš Leonardis, Elisa Ricci, Stefan Roth, Olga Russakovsky, Torsten S...

    2025

  7. [15]

    Jeonghwan Kim, Jisoo Kim, Jeonghyeon Na, and Hanbyul Joo. 2024. Parahome: Parameterizing everyday home activities towards 3d generative modeling of human-object interactions. arXiv preprint arXiv:2401.10232 (2024)

    2024 arXiv

  8. [16]

    Jiefeng Li, Siyuan Bian, Chao Xu, Zhicun Chen, Lixin Yang, and Cewu Lu. 2025. HybrIK-X: Hybrid Analytical-Neural Inverse Kinematics for Whole-Body Mesh Recovery. IEEE Trans. Pattern Anal. Mach. Intell. 47, 4 (Jan. 2025), 2754–2769

    2025

  9. [17]

    Jiaman Li, Jiajun Wu, and C Karen Liu. 2023. Object motion guided human motion synthesis. ACM Transactions on Graphics (TOG) 42, 6 (2023), 1–11

    2023

  10. [18]

    Jing Lin, Ailing Zeng, Shunlin Lu, Yuanhao Cai, Ruimao Zhang, Haoqian Wang, and Lei Zhang. 2023. Motion-x: A large-scale 3d expressive whole-body human motion dataset. Advances in Neural Information Processing Systems 36 (2023), 25268–25280

    2023

  11. [19]

    Haiyang Liu, Zihao Zhu, Giorgio Becherini, Yichen Peng, Mingyang Su, You Zhou, Xuefei Zhe, Naoya Iwamoto, Bo Zheng, and Michael J Black. 2024. EMAGE: Towards unified holistic co-speech gesture generation via expressive masked audio gesture modeling. In Proceedings of the IEEE/...

    2024

  12. [20]

    Shunlin Lu, Ling-Hao Chen, Ailing Zeng, Jing Lin, Ruimao Zhang, Lei Zhang, and Heung-Yeung Shum. 2023. Humantomato: Text-aligned whole-body motion generation. arXiv preprint arXiv:2310.12978 (2023)

    2023 arXiv

  13. [21]

    Xintao Lv, Liang Xu, Yichao Yan, Xin Jin, Congsheng Xu, Shuwen Wu, Yifan Liu, Lincheng Li, Mengxiao Bi, Wenjun Zeng, et al. 2024. HIMO: A New Benchmark for Full-Body Human Interacting with Multiple Objects. In European Conference on Computer Vision. Springer, 300–318

    2024

  14. [22]

    Gyeongsik Moon, Takaaki Shiratori, and Shunsuke Saito. 2024. Expressive whole- body 3D gaussian avatar. In European Conference on Computer Vision . Springer, 19–35

    2024

  15. [23]

    Noitom. [n. d.]. Noitom PN Hybrid VTS System. https://noitom.com/

  16. [24]

    Georgios Pavlakos, Vasileios Choutas, Nima Ghorbani, Timo Bolkart, Ahmed AA Osman, Dimitrios Tzionas, and Michael J Black. 2019. Expressive body capture: 3d hands, face, and body from a single image. In Proceedings of the IEEE/CVF conference on computer vision and pattern reco...

    2019

  17. [25]

    Matthias Plappert, Christian Mandery, and Tamim Asfour. 2016. The kit motion- language dataset. Big data 4, 4 (2016), 236–252

    2016

  18. [26]

    Manolis Savva, Angel X Chang, Pat Hanrahan, Matthew Fisher, and Matthias Nießner. 2016. Pigraphs: learning interaction snapshots from observations. ACM Transactions On Graphics (TOG) 35, 4 (2016), 1–12

    2016

  19. [27]

    Omid Taheri, Vasileios Choutas, Michael J Black, and Dimitrios Tzionas. 2022. Goal: Generating 4d whole-body motion for hand-object grasping. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition . 13263– 13273

    2022

  20. [28]

    Omid Taheri, Nima Ghorbani, Michael J Black, and Dimitrios Tzionas. 2020. GRAB: A dataset of whole-body human grasping of objects. In Computer Vision– ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceed- ings, Part IV 16 . Springer, 581–600

    2020

  21. [29]

    Guy Tevet, Sigal Raab, Brian Gordon, Yonatan Shafir, Daniel Cohen-Or, and Amit H Bermano. 2022. Human motion diffusion model. arXiv preprint arXiv:2209.14916 (2022)

    2022 arXiv

  22. [30]

    Tripo3d. [n. d.]. Generate 3D model Powered by AI in One Clip, within Seconds. https://www.tripo3d.ai/

  23. [31]

    Yuliang Xiu, Jinlong Yang, Xu Cao, Dimitrios Tzionas, and Michael J Black. 2023. Econ: Explicit clothed humans optimized via normal integration. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition . 512–523

    2023

  24. [32]

    Liang Xu, Xintao Lv, Yichao Yan, Xin Jin, Shuwen Wu, Congsheng Xu, Yifan Liu, Yizhou Zhou, Fengyun Rao, Xingdong Sheng, et al. 2024. Inter-x: Towards versatile human-human interaction analysis. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognitio...

    2024

  25. [33]

    Mengqing Xue, Yifei Liu, Ling Guo, Shaoli Huang, and Changxing Ding. 2025. Guiding Human-Object Interactions with Rich Geometry and Relations. arXiv preprint arXiv:2503.20172 (2025)

    2025 arXiv

  26. [34]

    Ling-An Zeng, Guohong Huang, Yi-Lin Wei, Shengbo Gu, Yu-Ming Tang, Jingke Meng, and Wei-Shi Zheng. 2025. ChainHOI: Joint-based Kinematic Chain Model- ing for Human-Object Interaction Generation. arXiv preprint arXiv:2503.13130 (2025)

    2025 arXiv

  27. [35]

    Juze Zhang, Jingyan Zhang, Zining Song, Zhanhe Shi, Chengfeng Zhao, Ye Shi, Jingyi Yu, Lan Xu, and Jingya Wang. 2024. HOI-Mˆ 3: Capture Multiple Humans and Objects Interaction within Contextual Environment. In Proceedings of the IEEE/CVF Conference on Computer Vision and Patte...

    2024

  28. [36]

    Jianrong Zhang, Yangsong Zhang, Xiaodong Cun, Yong Zhang, Hongwei Zhao, Hongtao Lu, Xi Shen, and Ying Shan. 2023. Generating human motion from textual descriptions with discrete representations. In Proceedings of the IEEE/CVF conference on computer vision and pattern recogniti...

    2023

  29. [37]

    Xiaohan Zhang, Bharat Lal Bhatnagar, Sebastian Starke, Vladimir Guzov, and Gerard Pons-Moll. 2022. Couch: Towards controllable human-chair interactions. In European Conference on Computer Vision . Springer, 518–535

    2022

  30. [38]

    Xiaohan Zhang, Bharat Lal Bhatnagar, Sebastian Starke, Ilya Petrov, Vladimir Guzov, Helisa Dhamo, Eduardo Pérez-Pellitero, and Gerard Pons-Moll. 2024. Force: Dataset and method for intuitive physics guided human-object interaction. CoRR (2024)

    2024

  31. [39]

    Yuhong Zhang, Jing Lin, Ailing Zeng, Guanlin Wu, Shunlin Lu, Yurong Fu, Yuanhao Cai, Ruimao Zhang, Haoqian Wang, and Lei Zhang. 2025. Motion-X++: A Large-Scale Multimodal 3D Whole-body Human Motion Dataset.arXiv preprint arXiv:2501.05098 (2025)

    2025 arXiv

  32. [2022]

    In Pro- ceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

    Generating Diverse and Natural 3D Human Motions From Text. In Pro- ceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). 5152–5161

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.