Pith. sign in

REVIEW 5 major objections 6 minor 49 references

Estimating Uncertainty in Galaxy Morphology Classification

T0 review · 5 major / 6 minor · reviewed 2026-08-14 · deepseek-v4-flash

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

Pith's one-line read This paper claims that uncertainty in galaxy morphology classification splits into four source types, and that a lightweight post-hoc estimator (UEGMC-P) can predict all four from frozen foundation-model representations without sampling.

desk verdict Useful taxonomy, stacked evaluation: the uncertainty gains are likely inflated and Eq (2) doesn't hold, but the idea merits a serious revision. read the letter →

arxiv 2608.08398 v1 pith:4RT5EAYQ submitted 2026-08-09 cs.AI astro-ph.IM

Kai Cheng , Ruoqi Wang , Qiong Luo This is my paper · ORCID
classification cs.AIastro-ph.IM
keywords uncertaintyquantificationgalaxymorphologyclassificationfoundationmodelspost-hocestimationepistemicaleatoricBrierscoreBayesianapproximation
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 claims that uncertainty in galaxy-morphology classification is not one number but four: uncertainty from limited model capacity, from imperfect or scarce image data, from mismatches introduced by supplementary reference materials such as schematic symbols and text, and from the intrinsic ambiguity of galaxies caught between evolutionary states. The authors build a Bayesian approximation baseline (UEGMC-B) that averages predictions over model sizes, pretraining datasets, multimodal input configurations, and a hand-chosen set of galaxy-evolution label transitions, then train lightweight per-type estimators (UEGMC-P) to reproduce those uncertainty scores directly from the hidden layers of a frozen foundation model, avoiding sampling. On two public galaxy-morphology benchmark datasets and two backbone architectures, UEGMC-P ranks the most uncertain samples better than prior post-hoc uncertainty methods on all four uncertainty types, and it adds fewer than six million parameters. If correct, this would let astronomers see why a classifier is unsure — whether to scale the model, gather cleaner data, fix reference standards, or defer an intrinsically ambiguous galaxy to a human expert — without retraining large models.

What carries the argument

The load-bearing object is the four-component decomposition of the predictive integral in Eq. (2), which the paper writes as $p(y|x,D)=\int_m\int_{x_M}\int_\theta p(y|m,\theta,x)\,p(m|\theta,x)\,p(x_M|x_g)\,p(\theta|D)\,d\theta\,dx_M\,dm$. Each uncertainty type is approximated by the Shannon entropy of a uniformly weighted average over a perturbation set (Eqs. (4)-(7)), and UEGMC-P replaces those averages with lightweight estimators. The key mechanism is feature fusion $\xi = \operatorname{Concat}(z^{(j)})$: concatenating hierarchical embeddings from the frozen backbone so each estimator sees both the image content and the model's predictive inclination, then regressing to Brier scores or Brier-score differences. The four resulting scores $\upsilon_\Theta$, $\upsilon_\mathcal{A}$, $\upsilon_{\mathcal{X}_M}$, and $\upsilon_\mathcal{R}$ are the outputs that let a user attribute misclassification risk to its source.

What would settle it

One direct check: replace the physically motivated label transitions in $\mathcal{R}$ with random alternative labels and rerun the morphology-boundary uncertainty evaluation; if the relative AUC does not drop, the physical-evolution content of $\mathcal{R}$ is not doing the work. A complementary check: on a set of galaxies with human disagreement (e.g., from a citizen-science decision tree), the boundary uncertainty scores should correlate with disagreement more strongly than plain model confidence does; if not, the fourth uncertainty type is just recalibrated confidence.

Watch

Extended reading notes

Core claim

The central claim is that the Bayesian predictive distribution for galaxy morphology can be decomposed into four physically motivated components — data aleatoric, model epistemic, modality aleatoric, and morphology boundary — and that each component can be estimated post-hoc. UEGMC-B realizes the decomposition as discrete entropy estimates by averaging over four perturbation sets: different model sizes for epistemic uncertainty, different pretraining corpora for data aleatoric uncertainty, different modality configurations (none, text, schematic symbols, or both) for modality aleatoric uncertainty, and label shifts along galaxy-evolution transitions for morphology boundary uncertainty. UEGMC-P then replaces these expensive averages with four small MLP estimators that map a concatenation of hierarchical embeddings from the frozen backbone to Brier-score-based targets. The paper reports that this sampling-free design yields the best or second-best relative AUC in nearly all of the compared settings, with the largest gains on morphology boundary uncertainty, which it interprets as evidence that the decomposition captures physical ambiguity rather than mere model confidence.

Load-bearing premise

The framework assumes that the small hand-chosen list of plausible galaxy-evolution label transitions, together with uniform weighting over model sizes, pretraining datasets, modality settings, and label perturbations, faithfully represents the true Bayesian uncertainty integrals.

Editorial extensions

If this is right

  • Model epistemic scores identify samples that a larger foundation model would correct, so a user can decide whether scaling up is worthwhile before paying for it.
  • Data aleatoric scores flag degraded or rare galaxy images that would benefit from more or better pretraining data, providing a metric for whether expensive additional pretraining will help.
  • Modality aleatoric scores measure when adding text or schematic reference standards hurts rather than helps, allowing a multimodal classifier to be corrected or re-aligned.
  • Morphology boundary scores can be used to defer intrinsically ambiguous galaxies to human astronomers, preserving effort for cases where machine labels are not trustworthy.
  • Because UEGMC-P is post-hoc, sampling-free, and adds under six million parameters, it can attach uncertainty estimates to an already-deployed frozen classifier at negligible cost.

Reading between the lines

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

  • Beyond the paper: the same four-way decomposition could be transplanted to other noisy astronomical classification tasks (transients, spectra, radio sources), but the categories would need re-grounding; the real test is whether the uncertainty types predict distinct intervention outcomes, not just higher ranking AUC.
  • Beyond the paper: the uniform weights over the perturbation sets are the paper's pragmatic choice; a probabilistic version that derives transition probabilities from galaxy-evolution models or inter-annotator agreement could either strengthen the boundary-uncertainty claim or reveal how much of it is an artifact of the chosen label list.
  • Beyond the paper: the evaluation protocol replaces uncertain samples with better predictions, which measures ranking quality; an alternative protocol comparing each uncertainty score to human annotation confidence would test whether the four numbers correspond to genuinely different physical and cognitive sources, as the paper argues.
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

5 major / 6 minor

Summary. This paper proposes UEGMC, a post-hoc uncertainty estimation framework for galaxy morphology classification. UEGMC defines four uncertainty types—model epistemic, data aleatoric, modality aleatoric, and morphology boundary—and offers two instantiations: UEGMC-B, a Bayesian entropy approximation over sets of models, pretraining datasets, modality configurations, and label perturbations; and UEGMC-P, lightweight MLP estimators that predict Brier-score-derived uncertainty values from frozen foundation-model features. The authors evaluate on Galaxy10 DECaLS and GalaxyMNIST with ViT and ConvNeXt backbones, reporting relative AUC gains when the most uncertain samples are refined or their predictions are replaced by better ones.

Significance. The motivation is sound and the application is timely: static foundation-model classifiers in astrophysics currently do not expose uncertainty, and a lightweight, sampling-free, post-hoc method would be practically valuable. The paper also makes a useful proposal in distinguishing multiple uncertainty sources, and it reports experiments across two public datasets and several UQ baselines, including an ablation over feature layers and a computational-efficiency measurement. However, as detailed below, the central validation is compromised by a formally incorrect derivation, a partly circular evaluation protocol, and missing statistics, so the claim of 'competitive performance' is not yet established.

major comments (5)
  1. [Methodology, Eq. (2)] Equation (2) is presented as an expansion of Eq. (1), but it is not a valid marginalization. Since x is defined as {x_g, x_M}, the left-hand side conditions on x_M while the right-hand side integrates over x_M; the left-hand side would at least have to be p(y|x_g,D) for the right-hand side to be a hierarchical marginalization. Moreover, p(m|θ,x) is introduced without a generative story or a likelihood, and the uniform substitutions in Eqs. (4)-(7) (p(θ|D)=1/|Θ|, etc.) are asserted rather than derived. Consequently, the morphology-boundary score computed in Eq. (7) and predicted by UEGMC-P is a sensitivity-to-label-perturbations measure under the hand-selected set R, not an established measure of intrinsic physical ambiguity. This point is load-bearing because the paper's central novelty is the physical interpretation of the four uncertainty types.
  2. [Experiments, Tables 1-4 and UEGMC-P training] The evaluation of UEGMC-P is circular for at least the first three uncertainty types. The estimators ρ are trained to regress Brier scores or Brier-score differences on the very model, data, and modality configurations used to define the AUC targets. For example, model epistemic uncertainty is defined as υ_Θ = ρ_T(ξ) − ρ_L(ξ), with ρ_T and ρ_L trained on Brier scores of tiny and large models, and Table 1 then measures how well υ_Θ ranks samples whose predictions improve under a large model. This measures the fit of the regressor to the target, not an independent uncertainty-estimation capability. The comparison is also unfair to baselines that do not use Brier-score training targets. The authors should evaluate UEGMC-P on uncertainty targets held out from training, or compare against baselines trained on the same target construction.
  3. [Table 4] Table 4 omits the UEGMC-B row, although the text states that 'UEGMC-P and its Bayesian baseline, UEGMC-B, consistently perform competitively' across the four uncertainty types. Without these numbers, the claim that UEGMC-B provides a valid baseline for the morphology-boundary type is unsubstantiated. The authors should report UEGMC-B in Table 4 or explain why it cannot be computed.
  4. [Experimental Setup and Tables 1-4] No variance information is reported for any experiment. All AUC values appear to be single runs, and no seeds, number of repeats, or significance tests are given. Some differences are small (e.g., Table 1, mAP rows differ by less than 0.5 AUC), so without error bars the claimed improvements over the next-best method are not statistically supported. The authors should provide repeated-run statistics or at least bootstrap confidence intervals for the headline comparisons.
  5. [Methodology, UEGMC-P, and Experimental Setup] The manuscript does not specify optimizer, learning rate, number of epochs, MLP width and depth, feature-layer fusion details, loss function, or the train/validation split used for the estimators. The reported inference speed and parameter count also lack hardware and batch-size context. Since UEGMC-P is the main algorithmic contribution, these details are needed for reproducibility.
minor comments (6)
  1. [Eq. (7)] Equation (7) is labeled H_XM but should be labeled H_R (or H_m) to distinguish it from the modality aleatoric entropy in Eq. (6); the current notation makes Eqs. (5)-(7) hard to follow.
  2. [Figure 5] Figure 5 reports Pearson correlations between UEGMC-P estimates and 'Oracle' measurements, but the text never defines the Oracle construction for each uncertainty type; without this definition the figure is difficult to interpret.
  3. [Experimental Setup, UQ methods under comparison] The MaxLogit baseline is used in Tables 1-3 but not in Table 4; if max logit cannot be defined for the morphology-boundary task, the authors should say so explicitly.
  4. [Methodology, Eq. (8)] The statement that the Brier score 'simultaneously measures reliability and uncertainty' is informal; the Brier score is a proper scoring rule for predictive distributions and is not itself a decomposition into reliability and uncertainty without further assumptions. The text should distinguish these concepts.
  5. [Experimental Setup, Evaluation Metrics] The evaluation protocol description says AUC is computed by substituting original estimates with a better one, but the exact substitution rule differs across Tables 1-4; a precise formal definition of the AUC protocol and its normalization would improve reproducibility.
  6. [Figure 6] The accuracy curves in Figure 6 have no error bars, and the y-axis ranges differ per panel, making cross-panel gains hard to compare; consider common scales or shaded confidence intervals.

Circularity Check

2 steps flagged · score 6.0 of 10

Equation (2) is a definitional decomposition, not an expansion of Eq. (1), and morphology-boundary uncertainty is defined as sensitivity to the hand-selected perturbation set R.

  1. self definitional [Methodology, Uncertainty Categorization for GMC, Eq. (2)]
    "we further expand Equation (1), the predictive distribution, into multiple integrals involving distinct probabilistic components as follows: p(y|x,D) = ∫m∫xM∫θ p(y|m,θ,x) p(m|θ,x) p(xM |xg) p(θ|D) dθdxM dm , (2) where xg is the galaxy image, xM is the supplementary multimodal data, x={xg, xM } is the input, and m is the latent variable for modeling the physical boundary. These four components systemically categorize the predictive uncertainty into four types."

    The left-hand side of Eq. (2) conditions on x={x_g,x_M}, while the right-hand side marginalizes over x_M, so the equation is not a valid expansion of Eq. (1). The four factors and their labels—model epistemic, data aleatoric, modality aleatoric, morphology boundary—are introduced by authorial choice, not derived from the Bayesian predictive distribution. The sentence 'These four components systemically categorize the predictive uncertainty into four types' is a definitional equivalence: the conclusion (four distinct uncertainty types) is identical to the premise (write four factors and name them). The paper's central categorical claim is therefore true by construction, not by derivation.

  2. self definitional [UEGMC-B, R definition and Eq. (7); Results, Table 4 discussion]
    "r ∈ R denotes a series of galaxy evolution phenomena since galaxies may exist in transitional evolving states and cause ambiguity. Therefore, we formulate a physics-consistent strategy to reveal morphological ambiguity by performing perturbations on the labels of samples based on the phenomena of galaxy evolution. ... Moreover, the results of the morphology boundary uncertainty in Table 4 demonstrate that UEGMC-P better captures the intrinsic evolutionary ambiguity of galaxies than previous methods."

    Morphology boundary uncertainty is defined in Eq. (7) as the entropy of predictions under label perturbations drawn from R, and R is constructed from the very evolutionary transitions that are said to cause the ambiguity. The later claim that the score 'captures the intrinsic evolutionary ambiguity' is a restatement of this definition: UEGMC-P measures sensitivity to the authors' hand-selected perturbation set, not an independently established physical-ambiguity target. If R is incomplete or the uniform weighting 1/|R| is unjustified, the score silently changes meaning, and the paper supplies no external validation separating R from the physical claim.

full rationale

The paper's empirical evaluation—comparing UEGMC-B and UEGMC-P AUC against baselines—is self-contained and not itself circular: it tests whether the operationalized uncertainty scores identify samples that improve under refinement. Self-citations, including Wang, Wang, and Luo 2025 for multimodal GMC, are not load-bearing. The circularity lies earlier, in the derivation. Eq. (2) is presented as an 'expansion' of the Bayesian predictive distribution, but it is actually the authors' definitional construction: the four uncertainty categories are created by labeling the chosen factors, and the left/right side mismatch over x_M shows the equation is not an identity. Consequently, the claim that UEGMC identifies four physically meaningful uncertainty types is equivalent to the decision to write four factors in Eq. (2). The morphology-boundary component is a sharper instance: R is the set of evolutionary label perturbations, so the 'intrinsic ambiguity' measured is by construction the sensitivity to that hand-set R. These are partial circularities in the central categorical claim; the predictive-regression and benchmark content retains independent empirical substance, keeping the score below 8.

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

The central method rests on hand-selected sets and uniform weighting assumptions. The only truly new entity is the morphology boundary uncertainty, but it is not independently validated. The estimators are trained on targets the authors define, which is a form of circularity.

free parameters (3)
  • Label perturbation set R = not specified (hand-constructed)
    The set of galaxy evolution transitions used to define morphology boundary uncertainty is chosen by the authors; the paper gives one example (Unbarred Loose Spiral to Unbarred Tight Spiral) but not the full set or a rule for inclusion.
  • Uniform weights in Bayesian approximation = 1/|Theta|, 1/|A|, 1/|XM|, 1/|R|
    The paper assumes uniform probabilities over model sizes, pretraining datasets, modality configurations, and label perturbations in Eqs. (4)-(7) without justification.
  • Estimator architecture and training hyperparameters = not reported
    The MLP estimators have under 6M parameters, but the depth, widths, learning rate, epochs, and loss weights are not given.
assumptions (4)
  • domain assumption The predictive distribution can be decomposed into four independent uncertainty components as in Eq. (2)
    The paper asserts the factorization p(y|x,D) equals an integral over p(y|m,theta,x) p(m|theta,x) p(xM|xg) p(theta|D), but the components are not shown to be independent or to correspond uniquely to the named uncertainty sources.
  • domain assumption The Brier score is a sufficient target for uncertainty estimation
    Brier score measures accuracy and calibration but does not capture all aspects of uncertainty; the paper uses it as the sole supervised target for training all estimators.
  • domain assumption Foundation models and their features are frozen and accessible
    UEGMC-P assumes a trained foundation model is available and that its hierarchical embeddings contain enough information to predict the uncertainty targets.
  • ad hoc to paper Label perturbation based on galaxy evolution captures morphological boundary ambiguity
    The construction of R from selected evolution phenomena is specific to this paper and not derived from an external taxonomy; the choice of which transitions count is arbitrary.
invented entities (3)
  • Morphology boundary latent variable m
    purpose: To model the intrinsic ambiguity of galaxies in transitional evolutionary states and define the morphology boundary uncertainty type.
    The latent variable m is introduced in Eq. (2) but is never estimated or validated; it serves only as a conceptual device to motivate the label perturbation heuristic.
  • Four uncertainty types (model epistemic, data aleatoric, modality aleatoric, morphology boundary)
    purpose: To categorize predictive uncertainty and provide separate estimators.
    These types are new constructs specific to this paper. They are operationalized through hand-picked sets (model sizes, pretraining data, modalities, label perturbations) rather than through an independent physical measurement.
  • Uncertainty Source Analysis Set
    purpose: Training set for UEGMC-P estimators, defined as a subset of M x Theta x A x XM x R.
    The selection of this set is not specified; it is a new training construct introduced for the method.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Estimating Uncertainty in Galaxy Morphology Classification." pith.science (2026). https://pith.science/paper/4RT5EAYQ

@misc{pith2026260808398,
  author       = {Pith},
  title        = {Pith review of: Estimating Uncertainty in Galaxy Morphology Classification},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/4RT5EAYQ}},
  note         = {Machine review of arXiv:2608.08398}
}
read the original abstract

Astronomers classify galaxy morphology to investigate cosmic evolution. While deep foundation models are increasingly utilized in Galaxy Morphology Classification (GMC), little work has been done on evaluating the uncertainty of GMC results. Uncertainty evaluation is important because astronomical data are inherently noisy due to instrumental and environmental limitations. Also, the continuous evolution of galaxies creates intrinsic morphological ambiguity. However, current foundation models operate as deterministic point estimators, failing to quantify the uncertainty. To overcome this limitation, we propose UEGMC, a post-hoc framework of Uncertainty Estimation for Galaxy Morphology Classification. It categorizes uncertainty in GMC into distinct types by model parameters, astronomical data, reference standards, or intrinsic physical ambiguities, thereby facilitating better classification. Our framework can directly predict uncertainties from representations extracted from the frozen backbones of foundation models, without computationally expensive sampling, therefore enabling fine-grained uncertainty evaluations. Our experimental results demonstrate that UEGMC provides competitive uncertainty quantification performance compared with previous methods.

Figures

Figures reproduced from arXiv: 2608.08398 by the authors.

Figure 1
Figure 1. Unaddressed uncertainty in foundation models and [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Illustration of galaxy morphology description and [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. An illustration of our proposed UEGMC-P framework. With the foundation model kept frozen, UEGMC-P undergoes [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Examples with different morphology boundary un [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 6
Figure 6. Figure 6: Comparison of accuracy performance gain while [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

49 extracted references · 45 canonical work pages

  1. [1]

    Systems without a Dark Halo , author=

    Encounters between Spherical Galaxies—I. Systems without a Dark Halo , author=. Monthly Notices of the Royal Astronomical Society , volume=. 2005 , publisher=

    work page 2005

  2. [2]

    Systems with a Dark Halo , author=

    Encounters between Spherical Galaxies—II. Systems with a Dark Halo , author=. Monthly Notices of the Royal Astronomical Society , volume=. 2005 , publisher=

    work page 2005

  3. [3]

    The Astrophysical Journal Letters , volume=

    The Role of Mergers in Early-Type Galaxy Evolution and Black Hole Growth , author=. The Astrophysical Journal Letters , volume=. 2010 , publisher=

    work page 2010

  4. [4]

    arXiv preprint arXiv:2502.09610 , year=

    Morphological Classification of Galaxies , author=. arXiv preprint arXiv:2502.09610 , year=

    arXiv

  5. [5]

    International Conference on Learning Representations , year =

    Alexey Dosovitskiy and Lucas Beyer and Alexander Kolesnikov and Dirk Weissenborn and Xiaohua Zhai and Thomas Unterthiner and Mostafa Dehghani and Matthias Minderer and Georg Heigold and Sylvain Gelly and Jakob Uszkoreit and Neil Houlsby , title =. International Conference on Learning Representations , year =

    work page

  6. [6]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Convnext V2: Co-Designing and Scaling Convnets with Masked Autoencoders , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    work page

  7. [7]

    Github Repository , howpublished =

    Leung Henry , title =. Github Repository , howpublished =. 2024 , publisher =

    work page 2024

  8. [8]

    Github Repository , howpublished =

    Walmsley, Mike , title =. Github Repository , howpublished =. 2022 , publisher =

    work page 2022

Show all 49 references
  1. [9]

    Monthly Notices of the Royal Astronomical Society , volume=

    Rotation-Invariant Convolutional Neural Networks for Galaxy Morphology Prediction , author=. Monthly Notices of the Royal Astronomical Society , volume=. 2015 , publisher=

    2015

  2. [10]

    Monthly Notices of the Royal Astronomical Society , volume=

    Galaxy Zoo 2: Detailed Morphological Classifications for 304 122 Galaxies from the Sloan Digital Sky Survey , author=. Monthly Notices of the Royal Astronomical Society , volume=. 2013 , publisher=

    2013

  3. [11]

    arXiv preprint arXiv:1506.02158 , year=

    Bayesian Convolutional Neural Networks with Bernoulli Approximate Variational Inference , author=. arXiv preprint arXiv:1506.02158 , year=

    arXiv

  4. [12]

    Journal of Machine Learning Research , volume=

    A Simple Approach to Improve Single-Model Deep Uncertainty Via Distance-Awareness , author=. Journal of Machine Learning Research , volume=

  5. [13]

    IEEE Transactions on Pattern Analysis and Machine Intelligence , year=

    A Review of Uncertainty Representation and Quantification in Neural Networks , author=. IEEE Transactions on Pattern Analysis and Machine Intelligence , year=

  6. [14]

    Proceedings of the ACM International Conference on Multimedia , pages=

    Galaxalign: Mimicking Citizen Scientists' Multimodal Guidance for Galaxy Morphology Analysis , author=. Proceedings of the ACM International Conference on Multimedia , pages=

  7. [15]

    NeurIPS 2024 Workshop Foundation Models for Science: Progress, Opportunities, and Challenges , year=

    Vision Foundation Models: Can They Be Applied to Astrophysics Data? , author=. NeurIPS 2024 Workshop Foundation Models for Science: Progress, Opportunities, and Challenges , year=

    2024

  8. [16]

    Journal of Open Source Software , volume=

    Zoobot: Adaptable Deep Learning Models for Galaxy Morphology , author=. Journal of Open Source Software , volume=. 2023 , publisher=

    2023

  9. [17]

    2024 , url=

    Siddharth Mishra-Sharma and YIDING SONG and Jesse Thaler , booktitle=. 2024 , url=

    2024

  10. [18]

    The Astrophysical Journal Letters , volume=

    Panic! at the Disks: First Rest-Frame Optical Observations of Galaxy Structure at z > 3 with Jwst in the Smacs 0723 Field , author=. The Astrophysical Journal Letters , volume=. 2022 , publisher=

    2022

  11. [19]

    Monthly Notices of the Royal Astronomical Society , volume=

    The Sami Galaxy Survey: Mass and Environment as Independent Drivers of Galaxy Dynamics , author=. Monthly Notices of the Royal Astronomical Society , volume=. 2021 , publisher=

    2021

  12. [20]

    Monthly Notices of the Royal Astronomical Society , volume=

    Statistics of Galaxy Mergers: Bridging the Gap between Theory and Observation , author=. Monthly Notices of the Royal Astronomical Society , volume=. 2022 , publisher=

    2022

  13. [21]

    Monthly Notices of the Royal Astronomical Society , pages=

    The Dynamics of Old Inner Galaxy Stars in Milky Way-Mass Galaxies Using Fire-2 Simulations , author=. Monthly Notices of the Royal Astronomical Society , pages=. 2026 , publisher=

    2026

  14. [22]

    Astronomy & Astrophysics , volume=

    A New Method for Age-Dating the Formation of Bars in Disc Galaxies-the Timer View on Ngc1433’s Old Bar and the Inside-out Growth of Its Nuclear Disc , author=. Astronomy & Astrophysics , volume=. 2023 , publisher=

    2023

  15. [23]

    International Conference on Machine Learning , pages=

    Decomposition of Uncertainty in Bayesian Deep Learning for Efficient and Risk-Sensitive Learning , author=. International Conference on Machine Learning , pages=

  16. [24]

    Advances in Neural Information Processing Systems , volume=

    Simple and Scalable Predictive Uncertainty Estimation Using Deep Ensembles , author=. Advances in Neural Information Processing Systems , volume=

  17. [25]

    Case Studies in Applied Bayesian Data Science: Cirm Jean-Morlet Chair, Fall 2018 , pages=

    Bayesian Neural Networks: An Introduction and Survey , author=. Case Studies in Applied Bayesian Data Science: Cirm Jean-Morlet Chair, Fall 2018 , pages=. 2020 , publisher=

    2018

  18. [26]

    International Conference on Machine Learning , pages=

    Uncertainty Estimation by Fisher Information-Based Evidential Deep Learning , author=. International Conference on Machine Learning , pages=

  19. [27]

    Chen, Mengyuan and Gao, Junyu and Xu, Changsheng , booktitle=

  20. [28]

    International Conference on Machine Learning , year=

    Gem-Fi: Gated Evidential Mixtures with Fisher Modulation , author=. International Conference on Machine Learning , year=

  21. [29]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Evidential Transformation Network: Turning Pretrained Models into Evidential Models for Post-Hoc Uncertainty Estimation , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  22. [30]

    Advances in Neural Information Processing Systems , volume=

    Pytorch: An Imperative Style, High-Performance Deep Learning Library , author=. Advances in Neural Information Processing Systems , volume=

  23. [31]

    Dinov2: Learning Robust Visual Features without Supervision , journal =

    Maxime Oquab and Timoth. Dinov2: Learning Robust Visual Features without Supervision , journal =. 2024 , url =

    2024

  24. [32]

    Monthly Notices of the Royal Astronomical Society , volume=

    Galaxy Zoo Decals: Detailed Visual Morphology Measurements from Volunteers and Deep Learning for 314 000 Galaxies , author=. Monthly Notices of the Royal Astronomical Society , volume=. 2022 , publisher=

    2022

  25. [33]

    International Conference on Machine Learning , pages=

    Learning Transferable Visual Models from Natural Language Supervision , author=. International Conference on Machine Learning , pages=

  26. [34]

    Nature Astronomy , volume=

    A Core in a Star-Forming Disc as Evidence of Inside-out Growth in the Early Universe , author=. Nature Astronomy , volume=. 2025 , publisher=

    2025

  27. [35]

    Nature Astronomy , volume=

    Constraints on the in situ and ex situ Stellar Masses in Nearby Galaxies Obtained with Artificial Intelligence , author=. Nature Astronomy , volume=. 2024 , publisher=

    2024

  28. [36]

    Nature Astronomy , volume=

    A Giant Disk Galaxy Two Billion Years after the Big Bang , author=. Nature Astronomy , volume=. 2025 , publisher=

    2025

  29. [37]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops , month =

    Abdullah, Ahmed and Ebert, Nikolas and Wasenm\"uller, Oliver , title =. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops , month =. 2026 , pages =

    2026

  30. [38]

    Information Fusion , volume=

    A Review of Uncertainty Quantification in Deep Learning: Techniques, Applications and Challenges , author=. Information Fusion , volume=. 2021 , publisher=

    2021

  31. [39]

    Monthly Notices of the Royal Astronomical Society , volume=

    Transfer Learning for Galaxy Morphology from One Survey to Another , author=. Monthly Notices of the Royal Astronomical Society , volume=. 2019 , publisher=

    2019

  32. [40]

    Advances in Neural Information Processing Systems , volume=

    What Uncertainties Do We Need in Bayesian Deep Learning for Computer Vision? , author=. Advances in Neural Information Processing Systems , volume=

  33. [41]

    RAS Techniques and Instruments , volume=

    Radio Galaxy Zoo: Towards Building the First Multipurpose Foundation Model for Radio Astronomy with Self-Supervised Learning , author=. RAS Techniques and Instruments , volume=. 2024 , publisher=

    2024

  34. [42]

    Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

    Acknowledging Focus Ambiguity in Visual Questions , author=. Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

  35. [43]

    Monthly Notices of the Royal Astronomical Society , volume=

    Practical Galaxy Morphology Tools from Deep Supervised Representation Learning , author=. Monthly Notices of the Royal Astronomical Society , volume=. 2022 , publisher=

    2022

  36. [44]

    Bias-Reduced Uncertainty Estimation for Deep Neural Classifiers , booktitle =

    Yonatan Geifman and Guy Uziel and Ran El. Bias-Reduced Uncertainty Estimation for Deep Neural Classifiers , booktitle =. 2019 , url =

    2019

  37. [45]

    Monthly Notices of the Royal Astronomical Society , volume=

    Imbalance Learning for Variable Star Classification , author=. Monthly Notices of the Royal Astronomical Society , volume=. 2020 , publisher=

    2020

  38. [46]

    Nature , volume=

    Strangulation as the Primary Mechanism for Shutting Down Star Formation in Galaxies , author=. Nature , volume=. 2015 , publisher=

    2015

  39. [47]

    Journal of Applied Meteorology and Climatology , volume=

    A New Vector Partition of the Probability Score , author=. Journal of Applied Meteorology and Climatology , volume=

  40. [48]

    Proceedings of the AAAI Conference on Artificial Intelligence , volume=

    Generalized Zero-Shot Learning for Point Cloud Segmentation with Evidence-Based Dynamic Calibration , author=. Proceedings of the AAAI Conference on Artificial Intelligence , volume=

  41. [49]

    Proceedings of the AAAI Conference on Artificial Intelligence , volume=

    Robust Uncertainty Quantification Using Conformalised Monte Carlo Prediction , author=. Proceedings of the AAAI Conference on Artificial Intelligence , volume=

Pith tools

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