Pith. sign in

REVIEW 3 major objections 21 references

Multi-scale weak lensing detection of galaxy clusters with source redshift tomography

T0 review · 3 major / 0 minor · reviewed 2026-07-14 · grok-4.5

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

Pith's one-line read A single source-redshift bin at z_min=0.4 finds as many weak-lensing clusters as multi-bin tomography on Euclid-like mocks.

desk verdict Honest pipeline result for Euclid-like WL peaks: under their wavelet + z_s,min-cut setup, one bin at 0.4 matches multi-bin combos, but the null gain may be partly baked into how they union detections. read the letter →

arxiv 2603.10636 v1 pith:26AC57KB submitted 2026-03-11 astro-ph.CO

L. Chappuis , S. Pires , G.W. Pratt , G. Leroy , A. Daurelle , C. Giocoli , C. Carbone This is my paper · ORCID
classification astro-ph.CO
keywords weaklensinggalaxyclusterssourceredshifttomographypeakdetectionEuclidmulti-scalewaveletphotometricredshifts
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

This paper asks whether splitting background galaxies into redshift bins and combining their weak-lensing peak maps can detect more galaxy clusters than one carefully chosen bin. Using a wavelet multi-scale peak finder on Euclid-like mocks, the authors build overlapping bins by raising the minimum source redshift and test every combination of one to four bins. They find that a single bin with minimum source redshift 0.4 matches the multi-bin combinations. Large-scale structure and photometric redshift errors shrink the expected gains, but the main limit is that false peaks accumulate when bins are merged, so purity falls at a fixed detection threshold. The result matters for Euclid-scale surveys: it suggests that simple single-bin maps already capture the detections that tomography was hoped to add, without the purity cost of combining bins.

What carries the argument

The z_s,min-cut technique: overlapping source-redshift bins formed by progressively raising the minimum source redshift, each producing a lensing map that is searched with a wavelet multi-scale peak detector, then combining the resulting peak catalogues.

What would settle it

Repeat the single-bin versus multi-bin comparison on real Euclid weak-lensing maps or an independent simulation suite with a different peak finder, and test whether multi-bin combinations recover significantly more true clusters at fixed purity than the z_s,min=0.4 single bin.

Watch

Extended reading notes

Core claim

On Euclid-like mocks, a single source-redshift bin with z_s,min=0.4 performs as well as any combination of up to four tomographic bins for weak-lensing cluster detection. Large-scale structure and photometric redshift errors reduce tomography’s potential gains, but the dominant limitation is the accumulation of spurious detections across bins, which lowers purity at a fixed detection threshold.

Load-bearing premise

That these mocks and this particular wavelet detector plus bin-combination rule are representative enough that single-bin equality will hold for real Euclid data and other peak finders.

Editorial extensions

If this is right

  • Single-bin weak-lensing cluster searches with a modest z_s,min cut can match multi-bin completeness and purity on Euclid-like data.
  • Analysis pipelines can avoid multi-bin peak combination without losing detections relative to the methods tested here.
  • Photometric redshift errors and large-scale structure already erase most of the theoretical tomographic gain before combination noise is counted.
  • If multi-bin methods are still used, spurious-peak accumulation must be controlled or purity will drop at fixed threshold.

Reading between the lines

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

  • Peak finders that suppress noise differently from the wavelet method might still gain from tomography where this detector does not.
  • Real Euclid selection functions more complex than the mocks could reopen a multi-bin advantage if the purity model is incomplete.
  • Earlier weak-lensing forecasts that assumed ideal redshift bins without false-positive accumulation may have overstated tomography’s cluster-detection gains.
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 / 0 minor

Summary. The manuscript studies whether source-redshift tomography can increase the number of galaxy-cluster detections obtained solely from weak-lensing peaks. Using a wavelet multi-scale peak finder, the authors construct overlapping Euclid-like source bins by raising a minimum source redshift cut (z_s,min) and combine the resulting peak catalogues via a z_s,min-cut procedure. On a progressive suite of mocks (isolated NFW haloes, N-body embedded clusters, true and photometric redshifts), they report that a single bin with z_s,min = 0.4 matches the performance of any combination of up to four bins. Large-scale structure and photo-z errors reduce tomographic gains, but the dominant limitation is said to be the accumulation of spurious detections across bins, which lowers purity at fixed detection threshold.

Significance. If the single-bin versus multi-bin equality is robust, the result is practically useful for Euclid-like surveys: it would justify simpler, non-tomographic peak pipelines without sacrificing catalogue size at fixed purity. The progressive mock design (NFW to N-body, true to photo-z) is a sound way to isolate contaminants. The claim is falsifiable and of direct interest to weak-lensing cluster cosmology. Significance is tempered by the fact that the result appears tied to one detector and one combination rule; a general statement about tomography would require broader validation.

major comments (3)
  1. The central claim that a single z_s,min=0.4 bin performs as well as multi-bin combinations rests on the z_s,min-cut combination rule (union of peaks at fixed detection threshold). Under a simple union, false positives accumulate nearly independently while true peaks largely overlap, so purity falls by construction. The abstract ranks spurious accumulation as the dominant limitation over LSS and photo-z; that ranking is only meaningful if purity is evaluated after combination without re-thresholding, peak matching, or multi-bin coincidence requirements. The manuscript should either (i) re-threshold or apply a coincidence cut after combination and re-compare single- vs multi-bin purity/completeness, or (ii) clearly restrict the claim to this specific combination rule rather than to tomography in general.
  2. Only one peak finder (the recently introduced wavelet multi-scale method) is used. The null gain of tomography may be method-specific. At minimum, the paper should discuss whether the same single-bin optimum is expected for other common peak finders (e.g., aperture mass, Gaussian-smoothed SNR maps) or provide a limited cross-check on one alternative, so that the result is not over-generalised to all weak-lensing peak detection.
  3. The abstract states that all combinations from one to four tomographic bins were considered and that z_s,min=0.4 is optimal, but does not specify how purity and completeness are defined after multi-bin combination, what detection threshold is held fixed, or how peaks are matched across bins and to true clusters. These definitions are load-bearing for the equality claim and must be stated explicitly (with tables or figures of purity vs completeness for single-bin vs best multi-bin) so the result can be reproduced and stress-tested.

Circularity Check

0 steps flagged · score 0.0 of 10

Empirical mock comparison of detection strategies; no derivation reduces to its inputs by construction.

full rationale

The paper is an experimental study of weak-lensing peak detection on progressively more realistic mocks (isolated NFW, N-body embedded haloes, true vs photometric redshifts, Euclid-like n(z)). Its central claim—that a single z_s,min=0.4 bin matches multi-bin z_s,min-cut combinations, with spurious accumulation as the dominant purity limiter—is an observed outcome under a fixed detection threshold and a stated combination rule, not a first-principles prediction or uniqueness theorem. Detection thresholds, bin edges, and the z_s,min-cut procedure are experimental knobs whose effects are measured against mocks; they are not redefined as the claimed result. There is no self-definitional loop, no fitted parameter renamed as a prediction of a closely related quantity, no load-bearing uniqueness imported from overlapping authors, and no ansatz smuggled in via self-citation. Methodological concerns about whether a simple union at fixed threshold makes purity drop somewhat by construction are correctness/generalizability issues, not circularity of a derivation chain. Score 0 is appropriate.

Assumptions & free parameters 3 free parameters · 4 assumptions · 1 invented entities

Abstract-only ledger. The work rests on standard weak-lensing and mock-cosmology assumptions plus the authors' operational definitions of bins, detector, and purity. No free cosmological parameters are fitted to claim a new constant; free choices are experimental (bin edges, number of bins, detection threshold). Invented entities are methodological constructs, not new physics.

free parameters (3)
  • z_s,min single-bin optimum (0.4) = 0.4
    Chosen/selected as the best single cut among progressive minimum-redshift cuts; performance equality claim is tied to this value.
  • Detection threshold (fixed purity comparison)
    Purity-at-fixed-threshold is the comparison metric; the numerical threshold is an experimental free choice not given in the abstract.
  • Number and edges of tomographic bins (1–4 overlapping) = 1 to 4 bins; z_s,max=3
    Bin construction by progressive z_s,min is a design choice that defines the multi-bin combinations being compared.
assumptions (4)
  • domain assumption Weak-lensing peak detections in multi-scale wavelet maps trace galaxy clusters sufficiently for catalog science in Euclid-like surveys.
    Background premise of the entire detection program stated in the opening of the abstract.
  • domain assumption Euclid-like source n(z) to z_s,max=3 and the progressive z_s,min overlapping bins adequately represent survey tomography.
    Source distribution and binning scheme used for all comparisons.
  • domain assumption Mock ladder from isolated NFW to N-body embedded haloes with true and photometric redshifts captures the main contaminants (LSS, photo-z).
    Progressive sophistication of mocks is the evidence base for attributing limitations.
  • ad hoc to paper Combining peak detections across redshift bins via the z_s,min-cut technique is a fair operationalization of 'tomographic information'.
    Specific combination method named in the abstract; other fusion rules might change the conclusion.
invented entities (1)
  • z_s,min-cut technique (multi-bin peak combination)
    purpose: Operational way to use source-redshift tomography by combining weak-lensing peak detections from maps with different minimum source redshifts.
    Named as the applied technique; methodological construct rather than new physics. Independent evidence outside this paper not established in the abstract.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Multi-scale weak lensing detection of galaxy clusters with source redshift tomography." pith.science (2026). https://pith.science/paper/26AC57KB

@misc{pith2026260310636,
  author       = {Pith},
  title        = {Pith review of: Multi-scale weak lensing detection of galaxy clusters with source redshift tomography},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/26AC57KB}},
  note         = {Machine review of arXiv:2603.10636}
}
abstract

Recently, a number of methods have emerged to detect galaxy clusters solely through their weak lensing signal. Using the recently-introduced wavelet multi-scale detection method, we focus here on the potential for the use of tomographic information of the source galaxies to increase the number of weak lensing detections. We apply the $z_{s,\mathrm{min}}$-cut technique, consisting of the combination of weak lensing peak detections emerging from lensing maps obtained using different source redshift bins, to mock data sets of progressively increasing sophistication. The source redshift distribution is chosen to be $Euclid$-like, with a maximum depth of $z_{s,\mathrm{max}}=3$, and overlapping tomographic redshift bins are constructed by progressively increasing the minimum source redshift $z_{s,\mathrm{min}}$. Considering all possible detection combinations from one to four tomographic bins, we find that a single source redshift bin, with $z_{s,\mathrm{min}}=0.4$, performs as well as the combination of multiple redshift bins. By running detections on synthetic clusters of varying complexity -- from isolated Navarro Frenk White haloes to haloes embedded in and formed within N-body cosmological simulations, and considering both true and photometric source redshifts -- we show that while large-scale structure contamination and photometric redshift errors reduce the potential gains of the tomographic approach, the dominant limitation is the accumulation of spurious detections across redshift bins, leading to decreased purity at a fixed detection threshold.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

21 extracted references · 1 canonical work pages

  1. [1]

    Ge-Ming Chiu. 2000. The odd-even turn model for adaptive routing.IEEE Transactions on parallel and distributed systems11, 7 (2000), 729–738

    2000

  2. [2]

    Dally and Hiromichi Aoki

    William J. Dally and Hiromichi Aoki. 1993. Deadlock-free adaptive routing in multicomputer networks using virtual channels.IEEE transactions on Parallel and Distributed Systems4, 4 (1993), 466–475

    1993

  3. [3]

    2004.Principles and practices of interconnection networks

    William James Dally and Brian Patrick Towles. 2004.Principles and practices of interconnection networks. Elsevier

    2004

  4. [4]

    José Duato. 1993. A new theory of deadlock-free adaptive routing in wormhole networks.IEEE transactions on parallel and distributed systems4, 12 (1993), 1320–1331

    1993

  5. [5]

    Binzhang Fu, Yinhe Han, Jun Ma, Huawei Li, and Xiaowei Li. 2011. An abacus turn model for time/space-efficient reconfigurable routing. InProceedings of the 38th annual international symposium on Computer architecture. 259–270

    2011

  6. [6]

    Christopher J Glass and Lionel M Ni. 1992. The turn model for adaptive routing. ACM SIGARCH Computer Architecture News20, 2 (1992), 278–287

    1992

  7. [7]

    Crispín Gomez, Francisco Gilabert, María Engracia Gomez, Pedro López, and José Duato. 2007. Deterministic versus adaptive routing in fat-trees. In2007 IEEE International Parallel and Distributed Processing Symposium. IEEE, 1–8

    2007

  8. [8]

    Paul Gratz, Boris Grot, and Stephen W Keckler. 2008. Regional congestion awareness for load balance in networks-on-chip. In2008 IEEE 14th International Symposium on High Performance Computer Architecture. IEEE, 203–214

    2008

Show all 21 references
  1. [9]

    Becker, George Michelogiannakis, James Balfour, Brian Towles, D

    Nan Jiang, Daniel U. Becker, George Michelogiannakis, James Balfour, Brian Towles, D. E. Shaw, John Kim, and William J. Dally. 2013. A detailed and flexible cycle-accurate Network-on-Chip simulator. In2013 IEEE International Symposium on Performance Analysis of Systems and Sof...

    2013 doi

  2. [10]

    Nan Jiang, John Kim, and William J Dally. 2009. Indirect adaptive routing on large scale interconnection networks. InProceedings of the 36th annual international symposium on Computer architecture. 220–231

    2009

  3. [11]

    Gwangsun Kim, Changhyun Kim, Jiyun Jeong, Mike Parker, and John Kim. 2016. Contention-based congestion management in large-scale networks. In2016 49th Annual IEEE/ACM International Symposium on Microarchitecture (MICRO). IEEE, 1–13

    2016

  4. [12]

    Yuliang Li, Rui Miao, Hongqiang Harry Liu, Yan Zhuang, Fei Feng, Lingbo Tang, Zheng Cao, Ming Zhang, Frank Kelly, Mohammad Alizadeh, et al. 2019. HPCC: High precision congestion control. InProceedings of the ACM special interest group on data communication. 44–58

    2019

  5. [13]

    Sean Lie. 2023. Cerebras architecture deep dive: First look inside the hard- ware/software co-design for deep learning.IEEE Micro43, 3 (2023), 18–30

    2023

  6. [14]

    Sheng Ma, Natalie Enright Jerger, and Zhiying Wang. 2011. DBAR: an efficient routing algorithm to support multiple concurrent applications in networks-on- chip. InProceedings of the 38th annual international symposium on Computer architecture. 413–424

    2011

  7. [15]

    Ted Nesson and S Lennart Johnsson. 1995. ROMM routing on mesh and torus net- works. InProceedings of the seventh annual ACM symposium on Parallel algorithms and architectures. 275–287

    1995

  8. [16]

    Andrea Pellegrini. 2021. Arm Neoverse N2: Arm’s 2 nd generation high perfor- mance infrastructure CPUs and system IPs. In2021 IEEE Hot Chips 33 Symposium (HCS). IEEE, 1–27

    2021

  9. [17]

    Daeho Seo, Akif Ali, Won-Taek Lim, and Nauman Rafique. 2005. Near-optimal worst-case throughput routing for two-dimensional mesh networks. In32nd International Symposium on Computer Architecture (ISCA’05). IEEE, 432–443

    2005

  10. [18]

    2005.Load-balanced routing in interconnection networks

    Arjun Singh. 2005.Load-balanced routing in interconnection networks. Stanford University

    2005

  11. [19]

    Herbert Sullivan and Theodore R Bashkow. 1977. A large scale, homogeneous, fully distributed parallel machine, I. InProceedings of the 4th annual symposium on Computer architecture. 105–117

    1977

  12. [20]

    Leslie G Valiant and Gordon J Brebner. 1981. Universal schemes for parallel communication. InProceedings of the thirteenth annual ACM symposium on Theory of computing. 263–277

    1981

  13. [21]

    Brian Zimmer, Rangharajan Venkatesan, Yakun Sophia Shao, Jason Clemons, Matthew Fojtik, Nan Jiang, Ben Keller, Alicia Klinefelter, Nathaniel Pinckney, Priyanka Raina, et al . 2019. A 0.11 pj/op, 0.32-128 tops, scalable multi-chip- module-based deep neural network accelerator w...

    2019

Pith tools

Reviewed July 14, 2026 · model on record in the stance chip above.