Pith. sign in

REVIEW 2 major objections 6 minor 30 references

Learning What's Real: Disentangling Signal and Measurement Artifacts in Multi-Sensor Data, with Applications to Astrophysics

T0 review · 2 major / 6 minor · reviewed 2026-07-12 · grok-4.5

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

Pith's one-line read Overlapping multi-telescope observations can train dual encoders that separate a galaxy’s true physical signal from instrument-specific distortions.

desk verdict Solid generative dual-encoder extension of multi-instrument disentanglement on real Legacy–HSC pairs; useful and mostly well-checked, with residual leakage and neighbor conditioning as the main soft spots. read the letter →

arxiv 2604.09787 v2 pith:INZWB3BG submitted 2026-04-10 astro-ph.IM astro-ph.GAcs.LG

Pablo Mercader-Perez , Carolina Cuesta-Lazaro , Daniel Muthukrishna , Jeroen Audenaert , V. Ashley Villar , David W. Hogg , Marc Huertas-Company , William T. Freeman This is my paper · ORCID
classification astro-ph.IMastro-ph.GAcs.LG
keywords disentangledrepresentationlearningmulti-instrumentastronomycounterfactualgenerationflowmatchinggalaxyimaginginstrumentsystematicsself-supervisedpretraining
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

Physical measurements always mix the signal of interest with sensor-dependent artifacts that confound inference and block clean multi-instrument analysis. This paper argues that when the same systems are observed by more than one instrument, those overlaps are enough to learn the split without an explicit instrument model. A dual-encoder architecture is trained only to reconstruct an unseen anchor image: one encoder sees the same source under a different instrument, the other sees different sources under the target instrument, and a flow-matching decoder combines the two latents. On roughly 100,000 cross-matched galaxy images from two major surveys, the resulting physics space supports property inference, rare-object search, and similarity retrieval that are not dominated by sensor pathologies, while the instrument space captures observing conditions and enables counterfactual cross-survey generation. The method is offered as a general self-supervised recipe for multi-sensor science: treat instrument differences as natural augmentations and force invariance through counterfactual reconstruction.

What carries the argument

Counterfactual generation with dual encoders: the physics encoder is given same-source, different-instrument views; the instrument encoder is given different-source, same-instrument views; a conditional flow-matching decoder reconstructs an anchor image neither encoder has seen, creating an architectural information bottleneck that forces the separation.

What would settle it

Replace the five nearest spatial neighbors with random galaxies from the same survey; if instrument-property prediction from the instrument latent collapses and physics-space outliers again become dominated by saturated pixels and noise bands, the shared-local-systematics premise has failed.

Watch

Extended reading notes

Core claim

A dual-encoder flow-matching model trained solely with a counterfactual reconstruction objective on overlapping multi-instrument observations learns representations that explicitly separate intrinsic physical signals from sensor-specific distortions and noise. Demonstrated on paired DESI Legacy and Hyper Suprime-Cam galaxy images, the physics latents align matched sources across surveys and yield morphologically meaningful outliers, while the instrument latents cluster by survey and predict local observing conditions, enabling unconfounded inference, instrument-independent search, and cross-instrument generation.

Load-bearing premise

Images of other galaxies lying within a few arcminutes of the target are assumed to share essentially the same instrument conditions, so they can teach the model the sensor’s behavior without leaking the target’s own physics.

Editorial extensions

If this is right

  • Physics embeddings support outlier detection of rare morphologies without instrument artifacts dominating the ranking.
  • Counterfactual generation can predict how a low-resolution survey object would appear under a deeper, higher-resolution instrument to prioritize follow-up.
  • Nearest-neighbor search in physics space retrieves physically similar objects independent of which survey supplied the query.
  • Analysis pipelines trained on one survey’s images can be applied, zero-shot, to counterfactual images generated from another survey.
  • The same training recipe—overlap pairs, dual encoders, counterfactual reconstruction—extends to other multi-sensor scientific domains.

Reading between the lines

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

  • Tightening or widening the instrument-encoder bottleneck is a practical dial for residual leakage of physical properties into the instrument space.
  • Outside dense overlapping footprints the method needs an unpaired extension; otherwise coverage remains limited to shared sky regions.
  • The same structural split could reduce instrument-driven false positives in multi-mission time-series archives where detector systematics currently dominate anomaly lists.
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

2 major / 6 minor

Summary. The paper proposes a dual-encoder architecture with a counterfactual flow-matching objective that uses overlapping multi-instrument observations to disentangle physics-invariant signals from instrument-specific artifacts. Training triplets consist of an anchor image (never seen by the encoders), a same-source different-instrument view for the physics encoder, and same-instrument different-source neighbors for the instrument encoder; the decoder reconstructs the anchor conditioned on both latents (Eq. 1, Fig. 1). On ~100k cross-matched Legacy–HSC galaxy images the method produces UMAP-separated latent spaces with pair alignment (Fig. 2), physics-space outliers free of instrumental pathologies (Fig. 3), asymmetric R^{2} probes against AION-1 and random-ResNet baselines (Fig. 4 / Table 5), instrument-independent nearest-neighbor retrieval (Fig. 5), and usable cross-instrument posterior samples (Figs. 6–7, MSE and pipeline-transfer checks). The authors present the approach as a general recipe for scientific multi-modal pretraining that treats sensor effects as natural augmentations.

Significance. If the disentanglement holds at the claimed level, the work supplies a practical architectural recipe for multi-instrument foundation models that avoids hand-engineered contrastive losses and explicit instrument forward models. The generative decoder additionally yields a data-driven noise model and counterfactual survey translation, both of direct utility for follow-up prioritization and pipeline transfer across surveys. Strengths that raise the contribution above a pure methods paper include the external-catalog probes (PROVABGS spectroscopy, shape catalogs, survey metadata), the cross-predict ResNet control that bounds shared spatial structure, the zero-shot ellipticity pipeline transfer (R^{2} 0.82 vs 0.81), and the public simplified codebase. Residual leakage is acknowledged and partially quantified; the result is therefore a solid, usable advance for astro-ph.IM and related multi-sensor domains rather than a complete solution to causal disentanglement.

major comments (2)
  1. §3 Data Triplets / Application: the instrument encoder is conditioned on up to five nearest spatial neighbors within 3 arcmin, under the premise that PSF/depth vary smoothly enough to share essentially the same instrument conditions. Residual R^{2} of the instrument latent for redshift, stellar mass and morphology (Fig. 4 / Table 5) and of the physics latent for some instrument quantities leave open how much private physics or large-scale structure co-varies with the neighbors. A controlled ablation that degrades neighbor quality (random same-survey images, larger angular radii, or shuffled sky positions) and re-measures the R^{2} asymmetry and pair alignment would quantify the load-bearing assumption; without it the claim of unconfounded physics representations remains only partially stress-tested.
  2. §4.1 Parameter inference and Appendix A.5: the cross-predict ResNet baseline shows that physics latents do not capture private instrument information above the shared spatial floor, which is a useful control. However, the instrument latent still retains non-trivial predictive power for physical properties that are only partially recoverable from imaging. The paper should either tighten the instrument bottleneck (capacity study already flagged as future work) or report a quantitative leakage metric (e.g., mutual information or residual R^{2} after regressing out the cross-predict baseline) so that the degree of residual confounding is stated rather than left as a qualitative caveat.
minor comments (6)
  1. Fig. 4 and Table 5: the main-text figure is hard to read at print scale; the appendix table is clearer. Consider moving the full numerical table into the main text or enlarging the figure panels.
  2. §4 Data Preprocessing: the decision to discard the HSC y-band and to upsample Legacy 31×31 to 48×48 via linear interpolation is stated but not ablated; a short note on sensitivity would help reproducibility.
  3. Eq. (1) and surrounding text: the notation for the variable-length sets of physics and instrument embeddings is dense; a short algorithmic box listing the sampling of j, k, neighbors and the attention conditioning would improve clarity.
  4. Fig. 3 caption and text: the full 160×160 cutouts are useful, but the main-text figure is small; consider a supplementary gallery of the top outliers with survey labels.
  5. Related work: the comparison to Audenaert et al. (contrastive dual-encoder) is clear; a one-sentence quantitative note on why the generative objective preserves high-resolution features that contrastive alignment discards would strengthen the positioning.
  6. Code availability: the simplified public codebase is welcome; stating the exact commit or DOI used for the reported numbers would aid exact reproduction.

Circularity Check

0 steps flagged · score 1.0 of 10

No load-bearing circularity; mild non-essential self-citation to related dual-encoder work by overlapping authors, while central claims rest on held-out reconstruction and external probes.

full rationale

The paper's core claim is architectural and empirical: a dual-encoder flow-matching model trained to reconstruct a held-out anchor (never seen by either encoder) from physics latents of the same source under a different instrument and instrument latents of different sources under the same instrument produces disentangled representations. Success is measured by UMAP separation (Fig. 2), outlier morphology free of artifacts (Fig. 3), R^{2} asymmetry on external spectroscopic/shape/instrument catalogs versus AION-1 and random-ResNet baselines (Fig. 4 / Table 5), nearest-neighbor retrieval, and generative MSE/posterior calibration on held-out pairs. None of these reduce by construction to a fitted free parameter of the claim, nor to a self-definitional identity. The only self-citation of note is to Audenaert et al. (2025) [11] (overlapping authors), which introduced a dual-encoder contrastive approach on simulated light curves; the present work explicitly distinguishes its generative objective, applies it to real multi-survey imaging, and does not invoke any uniqueness theorem or load-bearing result from that paper. The spatial-neighbor conditioning assumption is an empirical modeling choice, not a circular derivation. The paper is therefore self-contained against external benchmarks; score 1 only for the non-load-bearing related-work citation.

Assumptions & free parameters 4 free parameters · 5 assumptions · 1 invented entities

The central claim rests on standard generative modeling (flow matching), standard CNN/UNet components, and domain assumptions about multi-survey astronomy—not on new physical entities. Load-bearing free choices are architectural bottlenecks, neighbor selection for instrument context, and latent capacity; load-bearing domain assumptions are existence of cross-matched pairs and smooth spatial variation of systematics. No new particles or forces are postulated; 'physics latent' and 'instrument latent' are learned codes, not claimed fundamental quantities.

free parameters (4)
  • Instrument-neighbor count and angular radius = 5 neighbors, 3 arcmin
    Up to five nearest spatial neighbors within 3 arcminutes define the instrument conditioning set; this hand-chosen spatial prior is load-bearing for the instrument encoder.
  • Physics/instrument latent token geometry = (B, 4, 16)
    Each encoder maps to 4 tokens × 16 dims; capacity of this bottleneck controls how much information can leak between spaces and is chosen by architecture design, not derived.
  • UNet and training hyperparameters = see Appendix Tables 1–2
    Base channels 128, multipliers (1,2,4,4), LR 1e-4, 75k steps, batch 64, etc., are free design choices that affect reconstruction quality and thus the quality of the learned separation.
  • Image preprocessing (crop, upsample, arcsinh, zero-point) = 48×48, 4 channels, ZP 22.5
    Legacy 31×31→48×48 upsample and band alignment discard HSC y-band and force a common FOV; these choices define what 'same physics' means in pixels.
assumptions (5)
  • domain assumption Observation = f(Signal, Instrument) + Noise, with physics factors independent of measurement and instrument factors acting as confounders.
    Stated in Introduction as the causal generative story the architecture is designed to invert; not proved, standard measurement model.
  • ad hoc to paper Cross-matched multi-instrument observations of the same source provide a sufficient training signal to separate physics and instrument without an explicit forward model of either instrument.
    Core methodological premise of §3; success is empirical, not guaranteed by theory in the paper.
  • domain assumption Instrument systematics (PSF, depth, noise) vary smoothly on ~arcminute scales so nearby sources share instrument conditions with the anchor.
    Explicitly motivates the five-neighbor instrument conditioning in the Application section.
  • standard math Conditional flow matching can learn p(x_anchor | z_physics from other instruments, z_instrument from other sources) from finite paired data.
    Relies on Lipman et al. flow matching; treated as established generative modeling machinery.
  • ad hoc to paper An architectural information bottleneck (dual encoders never seeing the anchor) is sufficient to drive disentanglement without hand-engineered contrastive or β-VAE-style losses.
    Claimed in Related Work and Methods; supported by UMAP/probe results but not formally proved.
invented entities (1)
  • Physics latent space z_phy and instrument latent space z_ins as structural factors of variation independent evidence
    purpose: Serve as the two conditioning codes for counterfactual generation and as the representations used for inference, retrieval, and outlier detection.
    These are learned embeddings defined by the training objective, not new physical objects. independent_evidence is partial: external catalog probes and cross-survey pair alignment provide falsifiable checks, but the factors are not uniquely identified.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Learning What's Real: Disentangling Signal and Measurement Artifacts in Multi-Sensor Data, with Applications to Astrophysics." pith.science (2026). https://pith.science/paper/INZWB3BG

@misc{pith2026260409787,
  author       = {Pith},
  title        = {Pith review of: Learning What's Real: Disentangling Signal and Measurement Artifacts in Multi-Sensor Data, with Applications to Astrophysics},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/INZWB3BG}},
  note         = {Machine review of arXiv:2604.09787}
}
read the original abstract

Data collected from the physical world is always a combination of multiple sources: an underlying signal from the physical process of interest and a signal from measurement-dependent artifacts from the sensor or instrument. This secondary signal acts as a confounding factor, limiting our ability to extract information about the physics underlying the phenomena we observe. Furthermore, it complicates the combination of observations in heterogeneous or multi-instrument settings. We propose a deep learning framework that leverages overlapping observations, a dual-encoder architecture, and a counterfactual generation objective to disentangle these factors of variation. The resulting representations explicitly separate intrinsic signals from sensor-specific distortions and noise, and can be used for counterfactual view generation, parameter inference unconfounded by measurement distortions, and instrument-independent similarity search. We demonstrate the effectiveness of our approach on astrophysical galaxy images from the DESI Legacy Imaging Survey (Legacy) and the Hyper Suprime-Cam (HSC) Survey as a representative multi-instrument setting. This framework provides a general recipe for scientific and multi-modal self-supervised pretraining: construct training pairs from overlapping observations of the same physical system, treat sensor- or modality-specific effects as augmentations, and learn invariant representations through counterfactual generation.

Figures

Figures reproduced from arXiv: 2604.09787 by the authors.

Figure 1
Figure 1. Counterfactual Reconstruction Framework. The model learns to disentangle intrinsic galaxy properties from instrument systematics by reconstructing an anchor image via a dual-encoder architecture and conditional flow matching. Training uses data triplets consisting of: an anchor observation (signal s from instrument i), an instrument-augmented observation (same source s, a different instrument i ′ ), and a physics-au… view at source ↗
Figure 2
Figure 2. Multi-instrument Galaxy Reconstructions. Columns 1–2 (Input Conditioning): The target galaxy observed via an alternate instrument (input to the physics encoder) and a set of up to five different galaxies imaged by the target instrument (input to the instrument encoder). Column 3 (Ground Truth): The original anchor (target) image, withheld from the encoders. Columns 4–5 (Posterior Samples): Independent samples genera… view at source ↗
Figure 3
Figure 3. Left: Pixel-wise Z-score distribution of generated posterior samples relative to ground￾truth target images, across all pixels and held-out galaxies. For each pixel, x is the true value for a given pixel, xˆ is the predicted one, E[ˆx] is the posterior sample mean over pixels and galaxies and std(ˆx) is the posterior sample standard deviation. Both HSC-anchored and Legacy-anchored reconstructions closely approximate… view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Latent Space Disentanglement. UMAP projections of the physics (left) and instru￾ment (right) latent spaces. Orange and blue points represent HSC and Legacy images, respectively. Matched markers (△, ×, □, ◦) denote cross-survey pairs of the same galaxy. In the physics s…
Figure 5
Figure 5. Figure 5: Probing Latent Disentanglement via Downstream Regression. We report R2 scores for the prediction of physics-related properties (left) and instrumental properties (right) on four sets of representations: our physics encoder latents (blue), the instrument encoder latents…
Figure 6
Figure 6. Figure 6: Instrument-Invariant Nearest-Neighbor Search. We evaluate the disentangled latent spaces by performing nearest neighbor retrieval using paired HSC and Legacy observations as queries. Physics Space Retrieval: For both queries, the corresponding pair from the alternate s…
Figure 7
Figure 7. Figure 7: Spatial Structure Preservation. Power spectrum (left) and autocorrelation (right) per band for ground-truth and generated images. Each plot shows the true value (blue), the posterior mean from 32 generated samples (orange dashed line), and the 1σ posterior standard dev…

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

30 extracted references · 2 canonical work pages

  1. [1]

    J. B. Tenenbaum and W. T. Freeman. Separating style and content with bilinear models.Neural Computation, 12:1247 – 1283, 2000

    2000

  2. [2]

    beta-V AE: Learning basic visual concepts with a constrained variational framework

    Irina Higgins, Loic Matthey, Arka Pal, Christopher Burgess, Xavier Glorot, Matthew Botvinick, Shakir Mohamed, and Alexander Lerchner. beta-V AE: Learning basic visual concepts with a constrained variational framework. InInternational Conference on Learning Representations,

  3. [3]

    URLhttps://openreview.net/forum?id=Sy2fzU9gl

  4. [4]

    Toward causal representation learning.Proceedings of the IEEE, 109(5):612–634, 2021

    Bernhard Schölkopf, Francesco Locatello, Stefan Bauer, Nan Rosemary Ke, Nal Kalchbrenner, Anirudh Goyal, and Yoshua Bengio. Toward causal representation learning.Proceedings of the IEEE, 109(5):612–634, 2021. doi: 10.1109/JPROC.2021.3058954

    doi:10.1109/jproc.2021.3058954 2021

  5. [5]

    Evolutionary-scale prediction of atomic- level protein structure with a language model.Science, 379(6637):1123–1130, 2023

    Zeming Lin, Halil Akin, Roshan Rao, Brian Hie, Zhongkai Zhu, Wenting Lu, Nikita Smetanin, Robert Verkuil, Ori Kabeli, Yaniv Shmueli, Allan dos Santos Costa, Maryam Fazel-Zarandi, Tom Sercu, Salvatore Candido, and Alexander Rives. Evolutionary-scale prediction of atomic- level protein structure with a language model.Science, 379(6637):1123–1130, 2023. doi:...

    doi:10.1126/science.ade2574 2023

  6. [6]

    Large-scale chemical language representations capture molecular structure and properties

    Jerret Ross, Brian Belgodere, Vijil Chenthamarakshan, Inkit Padhi, Youssef Mroueh, and Payel Das. Large-scale chemical language representations capture molecular structure and properties. Nature Machine Intelligence, 4(12):1256–1264, 2022. URL https://arxiv.org/abs/2106. 09553

    2022

  7. [7]

    Satmae: Pre-training transformers for temporal and multi-spectral satellite imagery.Advances in Neural Information Processing Systems, 35: 197–211, 2022

    Yezhen Cong, Samar Khanna, Chenlin Meng, Patrick Liu, Erik Rozi, Yutong He, Marshall Burke, David Lobell, and Stefano Ermon. Satmae: Pre-training transformers for temporal and multi-spectral satellite imagery.Advances in Neural Information Processing Systems, 35: 197–211, 2022. URLhttps://arxiv.org/abs/2207.08051

    arXiv 2022

  8. [8]

    Mike Walmsley, Inigo Val Slijepcevic, Micah Bowles, and Anna M. M. Scaife. Towards galaxy foundation models with hybrid contrastive learning, 2022. URL https://arxiv.org/abs/ 2206.11927

    arXiv 2022

Show all 30 references
  1. [9]

    Aion-1: Omnimodal foundation model for astronomical sciences

    Liam Parker, Francois Lanusse, Jeff Shen, Ollie Liu, Tom Hehir, Leopoldo Sarra, Lucas Meyer, Micah Bowles, Sebastian Wagner-Carena, Helen Qu, Siavash Golkar, Alberto Bietti, Hatim Bourfoune, Nathan Casserau, Pierre Cornette, Keiya Hirashima, Geraud Krawezik, Ruben Ohana, Nicho...

    2025

  2. [10]

    The multimodal universe: Enabling large-scale machine learning with 100tb of astronomical scientific data, 2024

    The Multimodal Universe Collaboration. The multimodal universe: Enabling large-scale machine learning with 100tb of astronomical scientific data, 2024. URL https://arxiv. org/abs/2412.02527

    2024 arXiv

  3. [11]

    Siudek, M

    Euclid Collaboration, M. Siudek, M. Huertas-Company, M. Smith, G. Martinez-Solaeche, F. Lanusse, S. Ho, E. Angeloudi, P. A. C. Cunha, H. Domínguez Sánchez, M. Dunn, Y . Fu, P. Iglesias-Navarro, J. Junais, J. H. Knapen, B. Laloux, M. Mezcua, W. Roster, G. Stevens, J. Vega-Ferre...

    2025 doi arXiv

  4. [12]

    Gregory, David W

    Jeroen Audenaert, Daniel Muthukrishna, Paul F. Gregory, David W. Hogg, and V . Ashley Villar. Causal foundation models: Disentangling physics from instrument properties, 2025. URL https://arxiv.org/abs/2507.05333. 11

    2025 arXiv

  5. [13]

    Hogg, Dun Wang, Daniel Foreman-Mackey, Dominik Janzing, Carl-Johann Simon-Gabriel, and Jonas Peters

    Bernhard Schölkopf, David W. Hogg, Dun Wang, Daniel Foreman-Mackey, Dominik Janzing, Carl-Johann Simon-Gabriel, and Jonas Peters. Modeling confounding by half-sibling regression. Proceedings of the National Academy of Sciences, 113(27):7391–7398, 2016. doi: 10.1073/ pnas.15116...

    2016 doi

  6. [14]

    Hogg, Benjamin T

    Soichiro Hattori, Daniel Foreman-Mackey, David W. Hogg, Benjamin T. Montet, Ruth Angus, T. A. Pritchard, Jason L. Curtis, and Bernhard Schölkopf. The unpopular Package: A Data- driven Approach to Detrending TESS Full-frame Image Light Curves.The Astronomical Journal, 163(6):28...

    2022 doi

  7. [15]

    Yaron Lipman, Ricky T. Q. Chen, Heli Ben-Hamu, Maximilian Nickel, and Matt Le. Flow matching for generative modeling, 2023. URLhttps://arxiv.org/abs/2210.02747

    2023 arXiv

  8. [16]

    Schlegel, Dustin Lang, Robert Blum, Kaylan Burleigh, Xiaohui Fan, Joseph R

    Arjun Dey, David J. Schlegel, Dustin Lang, Robert Blum, Kaylan Burleigh, Xiaohui Fan, Joseph R. Findlay, Doug Finkbeiner, David Herrera, Stéphanie Juneau, Martin Landriau, Michael Levi, Ian McGreer, Aaron Meisner, Adam D. Myers, John Moustakas, Peter Nugent, Anna Patej, Edward...

    2019

  9. [19]

    Deep residual learning for image recognition, 2015

    Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Deep residual learning for image recognition, 2015. URLhttps://arxiv.org/abs/1512.03385

    2015 arXiv

  10. [20]

    Diffusers: State-of-the-art diffusion models

    Patrick von Platen, Suraj Patil, Anton Lozhkov, Pedro Cuenca, Nathan Lambert, Kashif Rasul, Mishig Davaadorj, Dhruv Nair, Sayak Paul, William Berman, Yiyi Xu, Steven Liu, and Thomas Wolf. Diffusers: State-of-the-art diffusion models. https://github.com/huggingface/ diffusers, 2022

    2022

  11. [21]

    Umap: Uniform manifold approximation and projection for dimension reduction, 2020

    Leland McInnes, John Healy, and James Melville. Umap: Uniform manifold approximation and projection for dimension reduction, 2020. URL https://arxiv.org/abs/1802.03426

    2020 arXiv

  12. [22]

    ChangHoon Hahn, K. J. Kwon, Rita Tojeiro, Malgorzata Siudek, Rebecca E. A. Canning, Mar Mezcua, Jeremy L. Tinker, David Brooks, Peter Doel, Kevin Fanning, Enrique Gaztañaga, Robert Kehoe, Martin Landriau, Aaron Meisner, John Moustakas, Claire Poppett, Gregory Tarle, Benjamin W...

    2023 doi

  13. [23]

    Saxe, Pang Wei Koh, Zhenghao Chen, Maneesh Bhand, Bipin Suresh, and Andrew Y

    Andrew M. Saxe, Pang Wei Koh, Zhenghao Chen, Maneesh Bhand, Bipin Suresh, and Andrew Y . Ng. On random weights and unsupervised feature learning. InProceedings of the 28th International Conference on International Conference on Machine Learning, ICML’11, page 1089–1096, Madiso...

    2011

  14. [24]

    Ricker, Joshua N

    George R. Ricker, Joshua N. Winn, Roland Vanderspek, David W. Latham, Gáspár Á. Bakos, Jacob L. Bean, Zachory K. Berta-Thompson, Timothy M. Brown, Lars Buchhave, Nathaniel R. Butler, R. Paul Butler, William J. Chaplin, David Charbonneau, Jørgen Christensen-Dalsgaard, Mark Clam...

    2015 doi

  15. [25]

    Borucki, David Koch, Gibor Basri, Natalie Batalha, Timothy Brown, Douglas Caldwell, John Caldwell, Jørgen Christensen-Dalsgaard, William D

    William J. Borucki, David Koch, Gibor Basri, Natalie Batalha, Timothy Brown, Douglas Caldwell, John Caldwell, Jørgen Christensen-Dalsgaard, William D. Cochran, Edna DeV ore, Edward W. Dunham, Andrea K. Dupree, Thomas N. Gautier, John C. Geary, Ronald Gilliland, Alan Gould, Ste...

    2010 doi

  16. [26]

    Koch, William J

    David G. Koch, William J. Borucki, Gibor Basri, Natalie M. Batalha, Timothy M. Brown, Douglas Caldwell, Jørgen Christensen-Dalsgaard, William D. Cochran, Edna DeV ore, Ed- ward W. Dunham, III Gautier, Thomas N., John C. Geary, Ronald L. Gilliland, Alan Gould, Jon Jenkins, Yoji...

    2010 doi

  17. [27]

    Bahcall, Steven Bickerton, James Bosch, Kevin Bundy, Peter L

    Hiroaki Aihara, Nobuo Arimoto, Robert Armstrong, Stéphane Arnouts, Neta A. Bahcall, Steven Bickerton, James Bosch, Kevin Bundy, Peter L. Capak, James H. H. Chan, Masashi Chiba, Jean Coupon, Eiichi Egami, Motohiro Enoki, Francois Finet, Hiroki Fujimori, Seiji Fujimoto, Hisanori...

  18. [28]

    doi: 10.1093/pasj/psx066

    doi

  19. [29]

    Schlegel, Dustin Lang, Robert Blum, Kaylan Burleigh, Xiaohui Fan, Joseph R

    Arjun Dey, David J. Schlegel, Dustin Lang, Robert Blum, Kaylan Burleigh, Xiaohui Fan, Joseph R. Findlay, Doug Finkbeiner, David Herrera, Stéphanie Juneau, Martin Landriau, Michael Levi, Ian McGreer, Aaron Meisner, Adam D. Myers, John Moustakas, Peter Nugent, Anna Patej, Edward...

    2019 doi

  20. [30]

    AstroCLIP: Cross-modal pre-training for astronomical foundation models.arXiv preprint arXiv:2310.03024, 2023

    François Lanusse, Liam Parker, Siavash Golkar, Miles Cranmer, Alberto Bietti, Michael Eick- enberg, Geraud Krawezik, Michael McCabe, Ruben Ohana, Mariel Pettee, Bruno Régaldo- Saint Blancard, Tiberiu Tesileanu, Kyunghyun Cho, and Shirley Ho. AstroCLIP: Cross-modal pre-training...

    2023 arXiv

  21. [31]

    Neural spline flows

    Conor Durkan, Artur Bekasov, Iain Murray, and George Papamakarios. Neural spline flows. In Advances in Neural Information Processing Systems, volume 32, 2019

    2019

  22. [32]

    Zuko: Normalizing flows in PyTorch, 2023

    François Rozet, Felix Andry, François Lanusse, and Gilles Louppe. Zuko: Normalizing flows in PyTorch, 2023. URLhttps://github.com/probabilists/zuko. 15 A Appendix A.1 Model Architecture We use a ResNet-18 [17] architecture for the encoders, plus a UNet2DConditionModel [18] for...

    2023

Pith tools

Reviewed July 12, 2026 · model on record in the stance chip above.