Pith. sign in

REVIEW 5 major objections 6 minor 1 cited by

Meta Curvature-Aware Minimization for Domain Generalization

T0 review · 5 major / 6 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read The paper claims that jointly minimizing the vanilla loss, the SAM surrogate gap, and the meta-learning surrogate gap—derived from a curvature metric—steers models to flatter minima and improves domain generalization on five benchmarks.

desk verdict A plausible empirical recipe for DG buried under a theory section that does not actually show it minimizes curvature. read the letter →

arxiv 2412.11542 v3 pith:AQP2NMJY submitted 2024-12-16 cs.CV cs.LG

Ziyang Chen , Yiwen Ye , Feilong Tang , Yongsheng Pan , Yong Xia This is my paper · ORCID
classification cs.CVcs.LG
keywords domaingeneralizationsharpness-awareminimizationcurvaturemetricflatminimameta-learningsurrogategapout-of-distributionHessianspectralnorm
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

Sharpness-aware minimization (SAM) guides models toward flat minima but can misfire because it measures sharpness through loss values. This paper argues that training should first minimize the loss and only later reduce curvature near a local minimum, and proposes a loss-value-independent curvature metric that becomes influential precisely as the model converges. From that metric it derives MeCAM, whose objective is the vanilla loss plus the SAM surrogate gap plus a meta-learning surrogate gap formed by MixStyle perturbation. The paper shows a PAC-Bayesian generalization bound, proves an $O(\log T/\sqrt{T})$ convergence rate, and reports an average 2.7% accuracy gain over ERM across PACS, VLCS, OfficeHome, TerraIncognita, and DomainNet, with top results on three of them.

What carries the argument

The curvature metric $\mathcal{C}(f(\theta)) = \|H(\theta)\| / (\|\nabla f(\theta)\|^2 + 1)$, which equals $\lambda_{\max}(H(\theta))$ at a local minimum when $H$ is positive semi-definite. The metric is small far from convergence because the gradient norm is large, and it grows as the gradient vanishes, so it selectively penalizes curvature near minima. The algorithm is then obtained by substituting the central-difference approximation $H \approx (f(\theta+h)+f(\theta-h)-2f(\theta))/h^2$ with $h=\delta$ along the SAM perturbation direction and absorbing $1/h^2$ into the weight $\alpha$, yielding the three-term objective in Eq. (11).

What would settle it

On a small network or synthetic loss landscape, compute the true largest eigenvalue of the Hessian at the endpoints found by MeCAM, ERM, and SAM, and compare it with the paper's curvature metric $\mathcal{C}$. If MeCAM endpoints do not have lower $\lambda_{\max}$ than SAM or ERM endpoints, or if $\mathcal{C}$ and $\lambda_{\max}$ diverge, the central claim that MeCAM minimizes spectral curvature would be refuted.

Watch

Extended reading notes

Core claim

The central discovery is a derivable training objective that simultaneously minimizes the regular training loss, the SAM surrogate gap $f(\theta+\delta)-f(\theta)$, and the meta-learning surrogate gap $m(\theta-\delta)-f(\theta)$. Using the identity $\lambda_{\max}(H(\theta)) = \|H(\theta)\| / (\|\nabla f(\theta)\|^2 + 1)$ at a local minimum, the paper treats the Hessian spectral norm as the sharpness measure, approximates the Hessian by central differences with step $\delta = \rho \nabla f(\theta)/(\|\nabla f(\theta)\|+\epsilon)$, and replaces $1/h^2$ with a hyperparameter $\alpha$. The meta term is justified by a first-order expansion equating $\alpha(f(\theta-\delta)-f(\theta))$ with $\beta(m(\theta-\delta)-f(\theta))$ under $\beta \le \alpha$. The paper claims MeCAM finds flatter minima than SAM and its variants and reports superior accuracy on five domain-generalization benchmarks.

Load-bearing premise

The whole curvature argument rests on treating a one-dimensional central difference along the SAM perturbation direction as the Hessian's spectral norm; if that scalar does not track the largest Hessian eigenvalue, the algorithm is not minimizing $\lambda_{\max}$ as claimed.

Editorial extensions

If this is right

  • MeCAM should reach flatter minima than SAM-style training, since it explicitly reduces curvature only near convergence.
  • The objective remains a first-order optimizer in practice, so it can be used like Adam without Hessian computation.
  • MeCAM integrates with existing DG methods such as SWAD, CORAL, and RSC without extra hyperparameter search and improves their average accuracy.
  • The PAC-Bayesian bound suggests that the surrogate-gap terms tighten the generalization upper bound, so minimizing them should transfer better to unseen domains.
  • The convergence rate matches first-order methods, so the extra curvature terms do not slow optimization asymptotically.

Reading between the lines

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

  • The curvature metric's denominator makes $\mathcal{C}$ scale-sensitive to the gradient norm, so a natural test is whether normalizing gradients before computing $\delta$ changes which minima MeCAM selects.
  • Because the central-difference step is tied to the SAM perturbation direction, the method measures curvature along one direction, not the full Hessian; a variant that averages several perturbation directions might track $\lambda_{\max}$ more faithfully.
  • The $\beta \le \alpha$ condition and the MixStyle-based meta term suggest that MeCAM could be combined with other domain-perturbation strategies, such as style transfer or adversarial augmentation, to strengthen the meta surrogate gap.
  • If the flatness claim is right, MeCAM should also improve performance in other out-of-distribution settings, such as corrupted or shifted test data in image classification, which the paper does not test.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

5 major / 6 minor

Summary. The paper proposes MeCAM (Meta Curvature-Aware Minimization), a sharpness-based training algorithm for domain generalization. Starting from a curvature metric C(f(θ)) = ||H(θ)||/(||∇f(θ)||²+1), the authors approximate the Hessian via central differences, replace the step-size with a tunable hyperparameter α, and arrive at an objective that minimizes the vanilla loss, a SAM surrogate gap f(θ+δ)−f(θ), and a meta-learning surrogate gap m(θ−δ)−f(θ), with Mixstyle used to construct the virtual meta-test domain. The paper claims a PAC-Bayesian generalization bound and an O(log T/√T) non-convex convergence rate, and reports experiments on PACS, VLCS, OfficeHome, TerraIncognita, and DomainNet under the DomainBed protocol, showing an average 2.7% gain over ERM and top results on several benchmarks.

Significance. If the theoretical derivation were sound, the paper would provide a principled connection between Hessian spectral-norm minimization and a practical SAM-plus-meta-learning update, which would be a useful contribution to sharpness-based domain generalization. The empirical evaluation is a genuine strength: five standard DG benchmarks, DomainBed protocol, three seeds, comparison with 22 DG methods and 6 sharpness-based methods, plus extensibility experiments. The paper also ships a concrete algorithm with hyperparameter search details. However, the central mechanistic claim—that MeCAM minimizes the Hessian spectral norm or λmax—is not established, because the derivation conflates a directional second derivative with a matrix norm and because Proposition 3.5 is proved circularly. The convergence proof does not match the algorithm's update rule. These issues are load-bearing for the paper's main narrative, even though the algorithm may still work as a heuristic combination of SAM and Mixstyle consistency.

major comments (5)
  1. [Theorem 3.2 and Eq. (5)] The central-difference identity is stated for the Hessian matrix, but the right-hand side is a scalar. For vector h, f(θ+h)+f(θ−h)−2f(θ) ≈ hᵀH(θ)h, which divided by ||h||² is a Rayleigh quotient, not the spectral norm ||H(θ)|| nor λmax. With h=δ=ρ∇f/||∇f||, this quantity is only the second derivative along the current gradient direction. A Hessian with a large eigenvalue in an orthogonal direction makes the surrogate small while λmax is large, so minimizing Eq. (7) is not shown to minimize λmax or ||H(θ)||. This conflation directly undermines the claim that MeCAM minimizes curvature around the local minimum.
  2. [Eqs. (6)-(7) and Table 5] Replacing 1/h² with a free hyperparameter α is not a step-size choice: h is a vector, so 1/h² is not defined, and the scalar α absorbs the unknown directional scale. Consequently, the term α(f(θ+δ)+f(θ−δ)−2f(θ)) is not an approximation of ||H(θ)||. This also affects Table 5, where C is computed as α|f(θ+δ)+f(θ−δ)−2f(θ)|/(||∇f(θ)||²+1); that quantity is not the metric C(f(θ)) defined in Eq. (2). The absolute value in Table 5 additionally differs from the signed term in Eq. (7), whose role when far from convergence is claimed to vanish.
  3. [Proposition 3.5 and Supplementary A.3] The proof of Proposition 3.5 is circular: Eq. (46) begins with 'By assuming this proposition holds' and then solves for β. The subsequent algebra, including the step f(θ)−m(θ)≈C′ and β≈α∇f/∇m+C′/δ, does not derive Eq. (10) from the definitions of f and m; it assumes the desired equivalence. Please either provide a direct derivation or explicitly present Eq. (10) as a heuristic assumption rather than a proposition.
  4. [Theorem 3.4 and Algorithm 1] The convergence proof uses the update dt = −ηt(γgt + αgsam_t + αgmeta_t) with γ = 1−2α (Eq. (17)), i.e., equal coefficients α on the SAM and meta terms. The actual update in Algorithm 1 is θt+1 ← θt − ηt((1−α−β)∇f(θt) + α∇f(θt+δt) + β∇m(θt−δt)). Unless β=α, which the reported hyperparameters in Table I do not satisfy (e.g., PACS α=0.1, β=0.1 only coincidentally; VLCS α=0.2, β=0.1), the proof's cancellation of the Hessian terms, which relies on equal coefficients for θ+δ and θ−δ, does not apply to the algorithm. The statement 'aligning our MeCAM with the convergence rate of first-order gradient methods' is therefore unsupported for the implemented method.
  5. [Proposition 3.3 and Supplementary A.1] The generalization bound is not rigorously derived. In Eq. (15), δ is a vector but δ² and δ∇f are written as scalars, and the Hessian matrix is replaced by a scalar central-difference term without justification. The inequality hat f(θ+δ) ≤ hat f(θ) + hat H(θ) is not a valid Taylor bound in multiple dimensions, and the identification of hat H(θ) with α(hat f(θ+δ)+hat f(θ−δ)−2hat f(θ)) repeats the conflation from Theorem 3.2. Either provide a correct derivation with explicit matrix-vector notation and a clear relation between the central-difference term and the Hessian, or state this as heuristic motivation.
minor comments (6)
  1. [Algorithm 1] Line 6 contains the typo 'Forwad' and should read 'Forward'.
  2. [Eq. (8)] The square-root expression contains a LaTeX artifact 'vuut' and is difficult to read; please reformat the equation.
  3. [Supplementary A.2.3, Eq. (39)] The L-smoothness expansion for f(θmeta_t) writes ⟨∇f(θsam_t), θmeta_{t+1} − θmeta_t⟩; the gradient should presumably be ∇f(θmeta_t). Please correct this typo.
  4. [References] Several reference entries contain stray numbers after the page ranges (e.g., [20], [88], [4]) that appear to be leftover formatting artifacts; these should be cleaned before publication.
  5. [Table 5 and Section 4.6] The caption says the table reports the curvature metric C, but the formula used is a directional central difference with α omitted. Please state the exact quantity reported and its relation to Eq. (2).
  6. [Abstract] The abstract says 'a flat minima' (singular) in two places; use 'a flat minimum' or 'flat minima' consistently.

Circularity Check

2 steps flagged · score 6.0 of 10

MeCAM's derivation is partially circular: Proposition 3.5's proof assumes the proposition, and Eqs. (5)-(7) identify a directional second derivative with the Hessian spectral norm, packaging the target objective into the approximation.

  1. self definitional [Supplementary A.3, Proof of Proposition 3.5 (and Eq. (10) in Section 3.3)]
    "By assuming this proposition holds, we have: α(f (θ − δ) − f (θ)) ≈ β(m(θ − δ) − f (θ))."

    Proposition 3.5 is exactly the statement α(f(θ−δ)−f(θ)) ≈ β(m(θ−δ)−f(θ)). The proof's first substantive step assumes that equality, and the rest of the proof only solves for the hyperparameter β (β ≈ α∇f/∇m + C′/δ) under the assumed equality. Therefore Eq. (10) is not established independently. Since Eq. (11) is obtained by 'combining both Eq. (7) and Eq. (10)', the meta-learning surrogate gap in the final MeCAM objective is posited rather than derived from the curvature metric.

  2. other [Section 3.3, Theorem 3.2 and Eqs. (5)-(7)]
    "Using the central difference method, we have H(θ) ≈ f (θ + h) + f (θ − h) − 2f (θ) h2 ... and utilize a hyperparameter α to replace 1 h2 ... minθ f (θ) + α(f (θ + δ) − f (θ)) + α(f (θ − δ) − f (θ))."

    For vector h, the scalar central difference equals hᵀH(θ)h/∥h∥², a directional second derivative, not the Hessian matrix H(θ) and not its spectral norm. With h=δ=ρ∇f/∥∇f∥, the quantity is the Rayleigh quotient along the current gradient only. Replacing 1/h² by α then turns Eq. (7) into exactly the two surrogate-gap terms that later appear in Eq. (11) as the SAM and meta-learning gaps. Thus the target objective is built into the approximation: the paper does not show that minimizing Eq. (7) controls λmax(H(θ)), which is the central mechanistic claim of MeCAM.

full rationale

The empirical evaluation is self-contained against external benchmarks (PACS, VLCS, OfficeHome, TerraIncognita, DomainNet) and the reported gains over ERM and other DG methods are not themselves circular. There is no load-bearing self-citation chain: the authors' own prior work is cited only in related work and is not used to justify the main construction. However, the derivation chain from the curvature metric C(f(θ)) to the final objective Eq. (11) contains two load-bearing reductions. First, Theorem 3.2 and Eqs. (5)-(7) treat a scalar directional second derivative as though it were the Hessian matrix or its spectral norm; with h=δ the resulting objective is the SAM surrogate term plus its mirror term, so Eq. (7) is the surrogate objective by construction rather than a proven λmax-minimizer. Second, Proposition 3.5, which supplies the meta-learning surrogate gap in Eq. (11), is proved by assuming the proposition itself in Supplementary A.3 and then solving for β. These steps mean the final algorithm is best understood as a proposed combination of SAM and Mixstyle-based meta-learning with per-dataset tuned weights α and β, rather than as a derived consequence of minimizing λmax. Because the central theoretical derivation is partly circular while the empirical content remains independent, a score of 6 is appropriate.

Assumptions & free parameters 3 free parameters · 5 assumptions · 0 invented entities

The central algorithm depends on several free hyperparameters (α, β, ρ) fitted per dataset, and on ad-hoc approximations in the derivations. No new physical entities are introduced.

free parameters (3)
  • α = 0.05-0.2 per dataset
    Replaces 1/h² in the central difference approximation; tuned on validation data per dataset.
  • β = 0.05-0.1 per dataset
    Balances the meta surrogate gap; introduced via Proposition 3.5 and tuned per dataset.
  • ρ = 0.01-0.2 per dataset
    SAM perturbation radius; tuned per dataset.
assumptions (5)
  • domain assumption H(θ) is positive semi-definite at local minima
    Used in Lemma 3.1 to equate the curvature metric with λmax. Not always guaranteed in non-convex deep learning.
  • standard math f is L-Lipschitz smooth
    Assumed in Theorem 3.4 convergence analysis and in the surrogate gap proofs.
  • standard math f is twice-differentiable
    Needed for the central difference approximation in Theorem 3.2.
  • ad hoc to paper f is monotonic and smooth within [θ+δ, θ−δ] when unconverged
    Invoked in Section 3.3 to claim the sum of the SAM and backward-difference gaps tends to 0 far from convergence.
  • ad hoc to paper f(θ) − m(θ) is a small constant C' (Prop 3.5)
    Assumed in the proof of Proposition 3.5 to derive the equivalence between the backward difference and the meta term; no empirical support.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Meta Curvature-Aware Minimization for Domain Generalization." pith.science (2026). https://pith.science/paper/AQP2NMJY

@misc{pith2026241211542,
  author       = {Pith},
  title        = {Pith review of: Meta Curvature-Aware Minimization for Domain Generalization},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/AQP2NMJY}},
  note         = {Machine review of arXiv:2412.11542}
}
read the original abstract

Domain generalization (DG) aims to enhance the ability of models trained on source domains to generalize effectively to unseen domains. Recently, Sharpness-Aware Minimization (SAM) has shown promise in this area by reducing the sharpness of the loss landscape to obtain more generalized models. However, SAM and its variants sometimes fail to guide the model toward a flat minimum, and their training processes exhibit limitations, hindering further improvements in model generalization. In this paper, we first propose an improved model training process aimed at encouraging the model to converge to a flat minima. To achieve this, we design a curvature metric that has a minimal effect when the model is far from convergence but becomes increasingly influential in indicating the curvature of the minima as the model approaches a local minimum. Then we derive a novel algorithm from this metric, called Meta Curvature-Aware Minimization (MeCAM), to minimize the curvature around the local minima. Specifically, the optimization objective of MeCAM simultaneously minimizes the regular training loss, the surrogate gap of SAM, and the surrogate gap of meta-learning. We provide theoretical analysis on MeCAM's generalization error and convergence rate, and demonstrate its superiority over existing DG methods through extensive experiments on five benchmark DG datasets, including PACS, VLCS, OfficeHome, TerraIncognita, and DomainNet. Code will be available on GitHub.

Figures

Figures reproduced from arXiv: 2412.11542 by the authors.

Figure 1
Figure 1. Accuracy of our MeCAM and existing DG methods on [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Comparison between SAM and the proposed curvature metric. (a) Illustration of a sharp local minimum at [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Accuracy of our MeCAM and other sharpness-based methods across various training iterations using each domain of PACS as [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Visualization of the loss landscapes of ERM, SAM, and [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

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

  1. Decoding the Flow: CauseMotion for Emotional Causality Analysis in Long-form Conversations

    cs.CL 2025-01 reject novelty 4.0 of 10

    CauseMotion reportedly improves LLM emotional-causality accuracy on long dialogues and on DiaASQ, but the headline gain of 8.7 percent is inconsistent with the reported numbers in Table II.

Reference graph

Works this paper leans on

88 extracted references · 71 canonical work pages · cited by 1 Pith paper

  1. [1]

    Survey on deep neural networks in speech and vision systems

    Mahbubul Alam, Manar D Samad, Lasitha Vidyaratne, Alexander Glandon, and Khan M Iftekharuddin. Survey on deep neural networks in speech and vision systems. Neuro- computing, 417:302–321, 2020. 1

    2020

  2. [2]

    A state-of-the-art survey on deep learning theory and architectures

    Md Zahangir Alom, Tarek M Taha, Chris Yakopcic, Ste- fan Westberg, Paheding Sidike, Mst Shamima Nasrin, Mah- mudul Hasan, Brian C Van Essen, Abdul AS Awwal, and Vijayan K Asari. A state-of-the-art survey on deep learning theory and architectures. electronics, 8(3):292, 2019. 1

    2019

  3. [3]

    Towards understanding sharpness-aware minimization

    Maksym Andriushchenko and Nicolas Flammarion. Towards understanding sharpness-aware minimization. In Int. Conf. Mach. Learn., pages 639–668. PMLR, 2022. 2

    2022

  4. [4]

    Invariant risk minimization

    Martin Arjovsky, L ´eon Bottou, Ishaan Gulrajani, and David Lopez-Paz. Invariant risk minimization. arXiv preprint arXiv:1907.02893, 2019. 5, 6, 4, 7, 8, 9

    arXiv 1907

  5. [5]

    Metareg: Towards domain generalization using meta- regularization

    Yogesh Balaji, Swami Sankaranarayanan, and Rama Chel- lappa. Metareg: Towards domain generalization using meta- regularization. Advances in neural information processing systems, 31, 2018. 3

    2018

  6. [6]

    Recognition in terra incognita

    Sara Beery, Grant Van Horn, and Pietro Perona. Recognition in terra incognita. In ECCV, pages 456–473, 2018. 5

    2018

  7. [7]

    Domain generalization by marginal transfer learning

    Gilles Blanchard, Aniket Anand Deshmukh, Urun Dogan, Gyemin Lee, and Clayton Scott. Domain generalization by marginal transfer learning. Journal of machine learning re- search, 22(2):1–55, 2021. 5, 6, 4, 7, 8, 9

    2021

  8. [8]

    Ex- ploiting domain-specific features to enhance domain gener- alization

    Manh-Ha Bui, Toan Tran, Anh Tran, and Dinh Phung. Ex- ploiting domain-specific features to enhance domain gener- alization. NeurIPS, 34:21189–21201, 2021. 5, 6

    2021

Show all 88 references
  1. [9]

    Swad: Domain generalization by seeking flat minima

    Junbum Cha, Sanghyuk Chun, Kyungjae Lee, Han-Cheol Cho, Seunghyun Park, Yunsung Lee, and Sungrae Park. Swad: Domain generalization by seeking flat minima. NeurIPS, 34:22405–22418, 2021. 6

    2021

  2. [10]

    Domain generalization by mutual-information regu- larization with pre-trained models

    Junbum Cha, Kyungjae Lee, Sungrae Park, and Sanghyuk Chun. Domain generalization by mutual-information regu- larization with pre-trained models. InECCV, pages 440–457. Springer, 2022. 5, 6

    2022

  3. [11]

    When vision transformers outperform resnets without pre-training or strong data augmentations

    Xiangning Chen, Cho-Jui Hsieh, and Boqing Gong. When vision transformers outperform resnets without pre-training or strong data augmentations. In ICLR, 2022. 3

    2022

  4. [12]

    Treasure in distribution: a domain randomiza- tion based multi-source domain generalization for 2d medi- cal image segmentation

    Ziyang Chen, Yongsheng Pan, Yiwen Ye, Hengfei Cui, and Yong Xia. Treasure in distribution: a domain randomiza- tion based multi-source domain generalization for 2d medi- cal image segmentation. In Int. Conf. Med. Image Comput. Comput.-Assist. Intervent., pages 89–99. Springer, 2023. 2

    2023

  5. [13]

    Madg: margin-based adversarial learning for domain gener- alization

    Aveen Dayal, Vimal KB, Linga Reddy Cenkeramaddi, C Mohan, Abhinav Kumar, and Vineeth N Balasubramanian. Madg: margin-based adversarial learning for domain gener- alization. NeurIPS, 36, 2024. 5, 6, 4, 7, 8, 9

    2024

  6. [14]

    Numerical meth- ods for unconstrained optimization and nonlinear equations

    John E Dennis Jr and Robert B Schnabel. Numerical meth- ods for unconstrained optimization and nonlinear equations. SIAM, 1996. 4

    1996

  7. [15]

    Sharp minima can generalize for deep nets

    Laurent Dinh, Razvan Pascanu, Samy Bengio, and Yoshua Bengio. Sharp minima can generalize for deep nets. In Int. Conf. Mach. Learn., pages 1019–1028. PMLR, 2017. 2

    2017

  8. [16]

    Efficient 9 sharpness-aware minimization for improved training of neu- ral networks

    Jiawei Du, Hanshu Yan, Jiashi Feng, Joey Tianyi Zhou, Lian- gli Zhen, Rick Siow Mong Goh, and Vincent Tan. Efficient 9 sharpness-aware minimization for improved training of neu- ral networks. In ICLR, 2021. 2

    2021

  9. [17]

    Unbiased met- ric learning: On the utilization of multiple datasets and web images for softening bias

    Chen Fang, Ye Xu, and Daniel N Rockmore. Unbiased met- ric learning: On the utilization of multiple datasets and web images for softening bias. In ICCV, pages 1657–1664, 2013. 5

    2013

  10. [18]

    Curvature measures

    Herbert Federer. Curvature measures. Transactions of the American Mathematical Society, 93(3):418–491, 1959. 3

    1959

  11. [19]

    Model- agnostic meta-learning for fast adaptation of deep networks

    Chelsea Finn, Pieter Abbeel, and Sergey Levine. Model- agnostic meta-learning for fast adaptation of deep networks. In Int. Conf. Mach. Learn., pages 1126–1135. PMLR, 2017. 2

    2017

  12. [20]

    Sharpness-aware minimization for efficiently improving generalization

    Pierre Foret, Ariel Kleiner, Hossein Mobahi, and Behnam Neyshabur. Sharpness-aware minimization for efficiently improving generalization. In ICLR, 2021. 1, 2, 3, 7, 8, 4, 5, 6, 9

    2021

  13. [21]

    Domain-adversarial training of neural networks

    Yaroslav Ganin, Evgeniya Ustinova, Hana Ajakan, Pas- cal Germain, Hugo Larochelle, Franc ¸ois Laviolette, Mario March, and Victor Lempitsky. Domain-adversarial training of neural networks. Journal of machine learning research , 17(59):1–35, 2016. 5, 6, 4, 7, 8, 9

    2016

  14. [22]

    Loss surfaces, mode connectivity, and fast ensembling of dnns

    Timur Garipov, Pavel Izmailov, Dmitrii Podoprikhin, Dmitry P Vetrov, and Andrew G Wilson. Loss surfaces, mode connectivity, and fast ensembling of dnns. NeurIPS, 31, 2018. 1, 2

    2018

  15. [23]

    In search of lost do- main generalization

    Ishaan Gulrajani and David Lopez-Paz. In search of lost do- main generalization. In ICLR, 2021. 5, 4

    2021

  16. [24]

    Deep residual learning for image recognition

    Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Deep residual learning for image recognition. In CVPR, pages 770–778, 2016. 5

    2016

  17. [25]

    Simplifying neu- ral nets by discovering flat minima

    Sepp Hochreiter and J ¨urgen Schmidhuber. Simplifying neu- ral nets by discovering flat minima. NeurIPS, 7, 1994. 1, 2

    1994

  18. [26]

    Self-challenging improves cross-domain generalization

    Zeyi Huang, Haohan Wang, Eric P Xing, and Dong Huang. Self-challenging improves cross-domain generalization. In ECCV, pages 124–140. Springer, 2020. 6

    2020

  19. [27]

    A survey of deep meta-learning

    Mike Huisman, Jan N Van Rijn, and Aske Plaat. A survey of deep meta-learning. Artificial Intelligence Review, 54(6): 4483–4541, 2021. 2

    2021

  20. [28]

    On the maximum hessian eigenvalue and generalization

    Simran Kaur, Jeremy Cohen, and Zachary Chase Lipton. On the maximum hessian eigenvalue and generalization. In I Can’t Believe It’s Not Better Workshop: Understanding Deep Learning Through Empirical Falsification , pages 51–

  21. [29]

    On large- batch training for deep learning: Generalization gap and sharp minima

    Nitish Shirish Keskar, Dheevatsa Mudigere, Jorge Nocedal, Mikhail Smelyanskiy, and Ping Tak Peter Tang. On large- batch training for deep learning: Generalization gap and sharp minima. In ICLR, 2016. 1, 2, 3

    2016

  22. [30]

    Domain generalization through meta-learning: a survey

    Arsham Gholamzadeh Khoee, Yinan Yu, and Robert Feldt. Domain generalization through meta-learning: a survey. Ar- tificial Intelligence Review, 57(10):285, 2024. 2

    2024

  23. [31]

    Selfreg: Self-supervised contrastive regu- larization for domain generalization

    Daehee Kim, Youngjun Yoo, Seunghyun Park, Jinkyu Kim, and Jaekoo Lee. Selfreg: Self-supervised contrastive regu- larization for domain generalization. In Proceedings of the IEEE/CVF International Conference on Computer Vision , pages 9619–9628, 2021. 5, 6

    2021

  24. [32]

    Adam: A method for stochastic opti- mization

    Diederik P Kingma. Adam: A method for stochastic opti- mization. arXiv preprint arXiv:1412.6980, 2014. 5, 7, 4, 6, 8, 9

    2014 arXiv

  25. [33]

    Out-of-distribution gener- alization via risk extrapolation (rex)

    David Krueger, Ethan Caballero, Joern-Henrik Jacobsen, Amy Zhang, Jonathan Binas, Dinghuai Zhang, Remi Le Priol, and Aaron Courville. Out-of-distribution gener- alization via risk extrapolation (rex). In Int. Conf. Mach. Learn., pages 5815–5826. PMLR, 2021. 5, 6, 4, 7, 8, 9

    2021

  26. [34]

    Fine-tuning can distort pre- trained features and underperform out-of-distribution

    Ananya Kumar, Aditi Raghunathan, Robbie Matthew Jones, Tengyu Ma, and Percy Liang. Fine-tuning can distort pre- trained features and underperform out-of-distribution. In ICLR, 2022. 5, 6

    2022

  27. [35]

    Asam: Adaptive sharpness-aware minimiza- tion for scale-invariant learning of deep neural networks

    Jungmin Kwon, Jeongseop Kim, Hyunseo Park, and In Kwon Choi. Asam: Adaptive sharpness-aware minimiza- tion for scale-invariant learning of deep neural networks. In ICLR, pages 5905–5914. PMLR, 2021. 2

    2021

  28. [36]

    Adaptive estimation of a quadratic functional by model selection

    Beatrice Laurent and Pascal Massart. Adaptive estimation of a quadratic functional by model selection. Annals of statis- tics, pages 1302–1338, 2000. 1

    2000

  29. [37]

    Gradient alignment for cross- domain face anti-spoofing

    Binh M Le and Simon S Woo. Gradient alignment for cross- domain face anti-spoofing. In CVPR, pages 188–199, 2024. 1, 2

    2024

  30. [38]

    Deeper, broader and artier domain generaliza- tion

    Da Li, Yongxin Yang, Yi-Zhe Song, and Timothy M Hospedales. Deeper, broader and artier domain generaliza- tion. In ICCV, pages 5542–5550, 2017. 5

    2017

  31. [39]

    Learning to generalize: Meta-learning for do- main generalization

    Da Li, Yongxin Yang, Yi-Zhe Song, and Timothy Hospedales. Learning to generalize: Meta-learning for do- main generalization. In AAAI, 2018. 3, 5, 6, 4, 7, 8, 9

    2018

  32. [40]

    Domain generalization with adversarial feature learning

    Haoliang Li, Sinno Jialin Pan, Shiqi Wang, and Alex C Kot. Domain generalization with adversarial feature learning. In CVPR, pages 5400–5409, 2018. 5, 6, 4, 7, 8, 9

    2018

  33. [41]

    Visualizing the loss landscape of neural nets

    Hao Li, Zheng Xu, Gavin Taylor, Christoph Studer, and Tom Goldstein. Visualizing the loss landscape of neural nets. NeurIPS, 31, 2018. 8

    2018

  34. [42]

    Friendly sharpness-aware minimization

    Tao Li, Pan Zhou, Zhengbao He, Xinwen Cheng, and Xiaolin Huang. Friendly sharpness-aware minimization. In CVPR, pages 5631–5640, 2024. 1, 7, 8, 4, 5, 6, 9

    2024

  35. [43]

    Domain generalization via conditional invari- ant representations

    Ya Li, Mingming Gong, Xinmei Tian, Tongliang Liu, and Dacheng Tao. Domain generalization via conditional invari- ant representations. In AAAI, 2018. 5, 6, 4, 7, 8, 9

    2018

  36. [44]

    Deep domain generaliza- tion via conditional invariant adversarial networks

    Ya Li, Xinmei Tian, Mingming Gong, Yajing Liu, Tongliang Liu, Kun Zhang, and Dacheng Tao. Deep domain generaliza- tion via conditional invariant adversarial networks. InECCV, pages 624–639, 2018. 2

    2018

  37. [45]

    Ms- net: multi-site network for improving prostate segmentation with heterogeneous mri data

    Quande Liu, Qi Dou, Lequan Yu, and Pheng Ann Heng. Ms- net: multi-site network for improving prostate segmentation with heterogeneous mri data. IEEE Trans. Med. Imaging, 39 (9):2713–2724, 2020. 1, 2

    2020

  38. [46]

    Towards efficient and scalable sharpness-aware minimization

    Yong Liu, Siqi Mai, Xiangning Chen, Cho-Jui Hsieh, and Yang You. Towards efficient and scalable sharpness-aware minimization. In CVPR, pages 12360–12370, 2022. 2

    2022

  39. [47]

    Decoupled weight decay regularization

    I Loshchilov. Decoupled weight decay regularization. arXiv preprint arXiv:1711.05101, 2017. 7, 4, 5, 6, 8, 9

    2017 arXiv

  40. [48]

    Pac-bayesian model averaging

    David A McAllester. Pac-bayesian model averaging. In Pro- ceedings of the twelfth annual conference on Computational learning theory, pages 164–170, 1999. 4 10

    1999

  41. [49]

    A survey on modeling and improving relia- bility of dnn algorithms and accelerators

    Sparsh Mittal. A survey on modeling and improving relia- bility of dnn algorithms and accelerators. Journal of Systems Architecture, 104:101689, 2020. 1

    2020

  42. [50]

    Domain generalization via invariant fea- ture representation

    Krikamol Muandet, David Balduzzi, and Bernhard Sch¨olkopf. Domain generalization via invariant fea- ture representation. In Int. Conf. Mach. Learn., pages 10–18. PMLR, 2013. 1, 2

    2013

  43. [51]

    Domain generaliza- tion via ensemble stacking for face presentation attack detec- tion

    Usman Muhammad, Jorma Laaksonen, Djamila Ro- maissa Beddiar, and Mourad Oussalah. Domain generaliza- tion via ensemble stacking for face presentation attack detec- tion. IJCV, pages 1–24, 2024. 2

    2024

  44. [52]

    Reducing domain gap by reduc- ing style bias

    Hyeonseob Nam, HyunJae Lee, Jongchan Park, Wonjun Yoon, and Donggeun Yoo. Reducing domain gap by reduc- ing style bias. In CVPR, pages 8690–8699, 2021. 5, 6, 4, 7, 8, 9

    2021

  45. [53]

    A method for solving the convex program- ming problem with convergence rate o (1/k2)

    Yurii Nesterov. A method for solving the convex program- ming problem with convergence rate o (1/k2). In Dokl akad nauk Sssr, page 543, 1983. 7, 4, 5, 6, 8, 9

    1983

  46. [54]

    Multi-view domain general- ization for visual recognition

    Li Niu, Wen Li, and Dong Xu. Multi-view domain general- ization for visual recognition. In ICCV, pages 4193–4201,

  47. [55]

    Numerical optimiza- tion

    Jorge Nocedal and Stephen J Wright. Numerical optimiza- tion. Springer, 1999. 4

    1999

  48. [56]

    Learning expla- nations that are hard to vary

    Giambattista Parascandolo, Alexander Neitz, Antonio Orvi- eto, Luigi Gresele, and Bernhard Sch¨olkopf. Learning expla- nations that are hard to vary. In ICLR. OpenReview, 2021. 4

    2021

  49. [57]

    Moment matching for multi-source domain adaptation

    Xingchao Peng, Qinxun Bai, Xide Xia, Zijun Huang, Kate Saenko, and Bo Wang. Moment matching for multi-source domain adaptation. In Proceedings of the IEEE/CVF inter- national conference on computer vision , pages 1406–1415,

  50. [58]

    Domain agnostic learning with disentangled rep- resentations

    Xingchao Peng, Zijun Huang, Ximeng Sun, and Kate Saenko. Domain agnostic learning with disentangled rep- resentations. In Int. Conf. Mach. Learn., pages 5102–5112. PMLR, 2019. 1, 2

    2019

  51. [59]

    Ef- ficient domain generalization via common-specific low-rank decomposition

    Vihari Piratla, Praneeth Netrapalli, and Sunita Sarawagi. Ef- ficient domain generalization via common-specific low-rank decomposition. In Int. Conf. Mach. Learn. , pages 7728–

  52. [60]

    Fishr: Invariant gradient variances for out-of-distribution generalization

    Alexandre Rame, Corentin Dancette, and Matthieu Cord. Fishr: Invariant gradient variances for out-of-distribution generalization. In Int. Conf. Mach. Learn. , pages 18347– 18377. PMLR, 2022. 5, 6, 4, 7, 8, 9

    2022

  53. [61]

    Imagenet large scale visual recognition challenge.IJCV, 115:211–252,

    Olga Russakovsky, Jia Deng, Hao Su, Jonathan Krause, San- jeev Satheesh, Sean Ma, Zhiheng Huang, Andrej Karpa- thy, Aditya Khosla, Michael Bernstein, et al. Imagenet large scale visual recognition challenge.IJCV, 115:211–252,

  54. [62]

    Distributionally robust neural networks

    Shiori Sagawa, Pang Wei Koh, Tatsunori B Hashimoto, and Percy Liang. Distributionally robust neural networks. In ICLR, 2019. 5, 6, 4, 7, 8, 9

    2019

  55. [63]

    Learning from synthetic data: Addressing domain shift for semantic segmentation

    Swami Sankaranarayanan, Yogesh Balaji, Arpit Jain, Ser Nam Lim, and Rama Chellappa. Learning from synthetic data: Addressing domain shift for semantic segmentation. In CVPR, pages 3752–3761, 2018. 1

    2018

  56. [64]

    Sand-mask: An enhanced gradient masking strategy for the discovery of invariances in domain generalization

    Soroosh Shahtalebi, Jean-Christophe Gagnon-Audet, Touraj Laleh, Mojtaba Faramarzi, Kartik Ahuja, and Irina Rish. Sand-mask: An enhanced gradient masking strategy for the discovery of invariances in domain generalization. arXiv preprint arXiv:2106.02266, 2021. 5, 6, 7, 8, 9

    2021 arXiv

  57. [65]

    Multi-adversarial discriminative deep domain generalization for face presentation attack detection

    Rui Shao, Xiangyuan Lan, Jiawei Li, and Pong C Yuen. Multi-adversarial discriminative deep domain generalization for face presentation attack detection. In CVPR, pages 10023–10031, 2019. 2

    2019

  58. [66]

    Gradient matching for domain generalization

    Yuge Shi, Jeffrey Seely, Philip Torr, N Siddharth, Awni Hannun, Nicolas Usunier, and Gabriel Synnaeve. Gradient matching for domain generalization. In ICLR, 2021. 5, 6

    2021

  59. [67]

    Deep coral: Correlation alignment for deep domain adaptation

    Baochen Sun and Kate Saenko. Deep coral: Correlation alignment for deep domain adaptation. InECCV, pages 443–

  60. [68]

    Rethinking domain generalization for face anti- spoofing: Separability and alignment

    Yiyou Sun, Yaojie Liu, Xiaoming Liu, Yixuan Li, and Wen- Sheng Chu. Rethinking domain generalization for face anti- spoofing: Separability and alignment. In CVPR, pages 24563–24574, 2023. 2

    2023

  61. [69]

    Statistical learning theory

    Vladimir Vapnik. Statistical learning theory. NY: Wiley, 2: 831–842, 1998. 5, 6, 8, 4, 7, 9

    1998

  62. [70]

    Deep hashing network for unsupervised domain adaptation

    Hemanth Venkateswara, Jose Eusebio, Shayok Chakraborty, and Sethuraman Panchanathan. Deep hashing network for unsupervised domain adaptation. In CVPR, pages 5018– 5027, 2017. 5

    2017

  63. [71]

    Cross-domain face presentation attack detection via multi- domain disentangled representation learning

    Guoqing Wang, Hu Han, Shiguang Shan, and Xilin Chen. Cross-domain face presentation attack detection via multi- domain disentangled representation learning. In CVPR, pages 6678–6687, 2020. 1, 2

    2020

  64. [72]

    Generalizing to unseen domains: A survey on do- main generalization

    Jindong Wang, Cuiling Lan, Chang Liu, Yidong Ouyang, Tao Qin, Wang Lu, Yiqiang Chen, Wenjun Zeng, and S Yu Philip. Generalizing to unseen domains: A survey on do- main generalization. IEEE Trans. Knowledge Data Eng., 35 (8):8052–8072, 2022. 2

    2022

  65. [73]

    Inter-class and inter- domain semantic augmentation for domain generalization

    Mengzhu Wang, Yuehua Liu, Jianlong Yuan, Shanshan Wang, Zhibin Wang, and Wei Wang. Inter-class and inter- domain semantic augmentation for domain generalization. IEEE TIP, 2024. 1, 2

    2024

  66. [74]

    Sharpness-aware gradient matching for domain generaliza- tion

    Pengfei Wang, Zhaoxiang Zhang, Zhen Lei, and Lei Zhang. Sharpness-aware gradient matching for domain generaliza- tion. In CVPR, pages 3769–3778, 2023. 1, 2, 3, 5, 6, 7, 8, 4, 9

    2023

  67. [75]

    CR-SAM: Cur- vature regularized sharpness-aware minimization

    Tao Wu, Tie Luo, and Donald C Wunsch II. CR-SAM: Cur- vature regularized sharpness-aware minimization. In AAAI, pages 6144–6152, 2024. 1, 2, 7, 8, 4, 5, 6, 9

    2024

  68. [76]

    Adversarial domain adaptation with domain mixup

    Minghao Xu, Jian Zhang, Bingbing Ni, Teng Li, Chengjie Wang, Qi Tian, and Wenjun Zhang. Adversarial domain adaptation with domain mixup. In AAAI, pages 6502–6509,

  69. [77]

    Pyhessian: Neural networks through the lens of the hessian

    Zhewei Yao, Amir Gholami, Kurt Keutzer, and Michael W Mahoney. Pyhessian: Neural networks through the lens of the hessian. In 2020 IEEE international conference on big data (Big data), pages 581–590. IEEE, 2020. 3

    2020

  70. [78]

    Adahessian: An adap- tive second order optimizer for machine learning

    Zhewei Yao, Amir Gholami, Sheng Shen, Mustafa Mustafa, Kurt Keutzer, and Michael Mahoney. Adahessian: An adap- tive second order optimizer for machine learning. In AAAI, pages 10665–10673, 2021. 7, 4, 5, 6, 8, 9 11

    2021

  71. [79]

    Adaptive methods for nonconvex optimization

    Manzil Zaheer, Sashank Reddi, Devendra Sachan, Satyen Kale, and Sanjiv Kumar. Adaptive methods for nonconvex optimization. NeurIPS, 31, 2018. 7, 4, 5, 6, 8, 9

    2018

  72. [80]

    Adaptive risk minimization: A meta-learning approach for tackling group shift

    Marvin Zhang, Henrik Marklund, Abhishek Gupta, Sergey Levine, and Chelsea Finn. Adaptive risk minimization: A meta-learning approach for tackling group shift. arXiv preprint arXiv:2007.02931, 8(9), 2020. 5, 6, 4, 7, 8, 9

    2007 arXiv

  73. [81]

    Flatness-aware minimization for domain generalization

    Xingxuan Zhang, Renzhe Xu, Han Yu, Yancheng Dong, Pengfei Tian, and Peng Cui. Flatness-aware minimization for domain generalization. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 5189– 5202, 2023. 7, 4, 5, 6, 8, 9

    2023

  74. [82]

    Gradient norm aware minimization seeks first-order flatness and improves generalization

    Xingxuan Zhang, Renzhe Xu, Han Yu, Hao Zou, and Peng Cui. Gradient norm aware minimization seeks first-order flatness and improves generalization. In CVPR, pages 20247–20257, 2023. 2, 3, 7, 1, 4, 5, 6, 8, 9

    2023

  75. [83]

    Learning to generate novel domains for domain gen- eralization

    Kaiyang Zhou, Yongxin Yang, Timothy Hospedales, and Tao Xiang. Learning to generate novel domains for domain gen- eralization. In ECCV, pages 561–578. Springer, 2020. 1, 2

    2020

  76. [84]

    Do- main generalization with mixstyle

    Kaiyang Zhou, Yongxin Yang, Yu Qiao, and Tao Xiang. Do- main generalization with mixstyle. In ICLR, 2021. 2, 4, 5, 6, 7, 8, 9

    2021

  77. [85]

    Do- main adaptive ensemble learning

    Kaiyang Zhou, Yongxin Yang, Yu Qiao, and Tao Xiang. Do- main adaptive ensemble learning. IEEE TIP, 30:8008–8018,

  78. [86]

    Domain generalization: A survey

    Kaiyang Zhou, Ziwei Liu, Yu Qiao, Tao Xiang, and Chen Change Loy. Domain generalization: A survey. IEEE TPAMI, 45(4):4396–4415, 2022. 2

    2022

  79. [87]

    Adabelief optimizer: Adapting stepsizes by the belief in observed gradients

    Juntang Zhuang, Tommy Tang, Yifan Ding, Sekhar C Tatikonda, Nicha Dvornek, Xenophon Papademetris, and James Duncan. Adabelief optimizer: Adapting stepsizes by the belief in observed gradients. NeurIPS, 33:18795–18806,

  80. [88]

    Surrogate gap minimization improves sharpness- aware training

    Juntang Zhuang, Boqing Gong, Liangzhe Yuan, Yin Cui, Hartwig Adam, Nicha C Dvornek, James s Duncan, Ting Liu, et al. Surrogate gap minimization improves sharpness- aware training. In ICLR, 2022. 1, 2, 3, 5 12 Meta Curvature-Aware Minimization for Domain Generalization Suppleme...

    2022

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.