Pith. sign in

REVIEW 4 major objections 7 minor 79 references

Adaptive Hardness-driven Augmentation and Alignment Strategies for Multi-Source Domain Adaptations

T0 review · 4 major / 7 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read A3MDA claims that hardness-adaptive augmentation and alignment, built from three progressive hardness measurements, outperforms existing multi-source domain adaptation methods on seven benchmarks.

desk verdict Genuine method contribution with matching ablations, but the reported gains are not yet verified: tuning on test tasks, no variance, and an R inconsistency in the text. read the letter →

arxiv 2501.01142 v2 pith:BPLZVZUC submitted 2025-01-02 cs.CV

Yang Yuxiang , Zeng Xinyi , Zeng Pinxian , Zu Chen , Yan Binyu , Zhou Jiliu , Wang Yan This is my paper · ORCID
classification cs.CV
keywords Multi-sourcedomainadaptationHardness-awarelearningAdaptivedataaugmentationWeightedmaximummeandiscrepancyPseudo-labelrefinementContrastiveUnsupervisedTransfer
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 multi-source domain adaptation (MDA) underperforms because it treats every sample alike, and that a hardness-driven training loop can fix three weaknesses at once: uncontrolled strong augmentation, coarse sample-level inter-domain alignment, and missing intra-domain alignment. The proposed A3MDA estimates each sample's difficulty with three progressive measurements and uses them to adapt augmentation intensity, to weight a class-aware maximum mean discrepancy (MMD) loss, and to select harder target samples for a pseudo-contrastive loss. On seven public benchmarks the method reports higher average classification accuracy than the compared MDA methods. If the claim holds, a single hardness signal can organize both data augmentation and domain alignment in MDA.

What carries the argument

The carrying mechanism is the three-stage Adaptive Hardness Measurement (AHM) hierarchy. Basic AHM gauges instantaneous uncertainty as the $\ell^2$ norm of the prediction vector with the true or predicted class contribution removed. Smooth AHM applies an exponential moving average with factor $\beta$ over the training history, stored in a hardness memory, to stabilize epoch-to-epoch fluctuation. Comparative AHM divides smoothed hardness by the batch or class-cluster sum, producing relative weights that avoid any single sample dominating a batch. These three measurements feed three utilization actions: adjusting augmentation intensity, weighting clustered MMD, and selecting hard target samples for the pseudo-contrastive matrix. Each action is guided by the rule that harder samples receive lighter perturbation but stronger alignment pressure.

What would settle it

Compare A3MDA's cluster losses (Eqs. (6)-(9)) in two versions: one using all target pseudo-labels and one using only pseudo-labels whose confidence exceeds $\tau=0.6$. If the filtered version matches or beats the full version on the seven benchmarks, the claim that low-confidence hard samples can be safely included in the alignment losses is falsified.

Watch

Extended reading notes

Core claim

The central claim is that replacing uniform, sample-agnostic treatment of source and target data with hardness-adaptive treatment improves MDA classification. A3MDA defines Basic AHM $\boldsymbol{\Omega}_b(\cdot)$ as the $\ell^2$ norm of the prediction vector with the true (source) or pseudo (target) class element set to zero; Smooth AHM $\mathbf{S}_b(\cdot)$ is an exponential moving average of Basic values stored in a hardness memory; Comparative AHM $\mathbf{H}_b^{[c]}(\cdot)$ normalizes smoothed hardness within a batch or within a class cluster. These values drive three actions: Eq. (4) mixes weak and strong augmentation according to hardness; Eqs. (6)-(7) turn MMD into a weighted, cluster-level inter-domain loss; Eqs. (8)-(9) build a pseudo-contrastive matrix from harder target samples, aligning the target feature space and correcting pseudo-labels. The paper reports that the full system attains the best average accuracy among compared methods on Office-31, Office-Home, DomainNet, ImageCLEF-DA, PACS, Digits-5, and Office-Caltech.

Load-bearing premise

The load-bearing premise is that the guessed labels assigned to every target image, including the low-confidence ones, are accurate enough to specify which image features should be pulled together and which should be pushed apart during alignment.

Editorial extensions

If this is right

  • If A3MDA is right, a single adaptive augmentation scheme can replace per-dataset augmentation search, since hardness smoothing automatically raises intensity as the model generalizes.
  • The weighted-clustered MMD should give finer inter-domain alignment than sample-level MMD, because class attributes decide which pairs are pulled together and hardness decides how strongly.
  • Selecting harder samples for the pseudo-contrastive matrix should suppress erroneous pseudo-labels and produce better-separated target clusters, as the pseudo-label trend analysis claims.
  • The hardness machinery should transfer to other discrepancy metrics such as L2, Wasserstein, and CORAL, with the reported gains on Office-Home.
  • Across the seven tested benchmarks, the full method reports the highest average accuracy among the compared MDA methods.

Reading between the lines

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

  • Editorial inference: the hardness rankings could be reused as a curriculum schedule, feeding easier samples to the model before harder ones, a use the paper does not evaluate.
  • Editorial inference: the pseudo-contrastive matrix behaves like a contrastive objective with pseudo-label-defined positives; combining it with confidence filtering on the selected hard samples may stabilize early training further.
  • Editorial inference: because the framework is metric-agnostic and augmentation-based, it may extend to semantic segmentation or cross-modal adaptation, where hardness can be computed from prediction entropy at little extra cost.
  • Editorial inference: the claim implies a testable ordering—hardness-weighted alignment should help most on benchmarks with high intra-domain noise—so accuracy gains could be checked against target-domain prediction entropy across datasets.
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

4 major / 7 minor

Summary. The paper proposes A3MDA, a hardness-driven framework for multi-source domain adaptation (MDA). Three progressive Adaptive Hardness Measurements (Basic, Smooth, and Comparative AHMs) are derived from the model's own predictions: Basic AHM scores instantaneous sample difficulty, Smooth AHM temporally stabilizes these scores via an exponential moving average stored in a hardness memory, and Comparative AHM renormalizes hardness within batches or within-class clusters. These measurements are then used in three ways: Smooth AHM intensities modulate the proportion of weak versus strong augmentation per sample; Comparative AHM values serve as sample weights in a weighted-clustered variant of MMD that aligns same-class source and target clusters while repelling different classes; and hardness rankings select the hardest target samples to build a pseudo-contrastive matrix that enforces intra-domain clustering. The combined objective is evaluated on seven benchmarks (Office-31, Office-Home, DomainNet, ImageCLEF-DA, PACS, Digits-5, Office-Caltech), with the central claim that A3MDA outperforms existing MDA methods in average classification accuracy.

Significance. The conceptual direction is timely and reasonably well executed: applying sample-specific hardness to jointly control augmentation intensity and cluster-level alignment is a sensible extension of prior hardness- and uncertainty-based MDA work, and the paper goes beyond qualitative hard-sample selection by using hardness values quantitatively in the loss. Strengths that deserve explicit credit include the breadth of the evaluation (seven benchmarks with consistent protocols), the component-wise ablation (Table VIII) showing that each of the three utilization mechanisms contributes, the extensions to other discrepancy metrics (Table X) and backbones (Table IX), the explicit computational-cost comparison (Table VII), and the targeted robustness experiment with corrupted pseudo-labels (Fig. 11(c)). However, the empirical support for the headline claim is not yet convincing: no variance information is provided, the reported margins over the second-best method are small (0.6 to 1.5 points across the seven tables), and hyperparameters were selected using tasks that later appear in the evaluation tables.

major comments (4)
  1. [Section IV.B (Implementation and Training Details)] The evaluation protocol does not currently secure the central claim that A3MDA 'outperforms other methods.' The paper states 'We maintain a fixed random seed of 10 over 3 runs and report the average results'; with a fixed seed, the three runs are not independent replicates, and no standard deviations or significance tests are reported anywhere in Tables I-VI. The margins over the second-best method are 0.9 (Office-31), 0.8 (Office-Home), 1.0 (DomainNet), 1.5 (ImageCLEF-DA), 1.07 (PACS), 0.6 (Digits-5), and 0.7 (Office-Caltech) percentage points, all of which are within the run-to-run variation typically observed for deep MDA methods. The authors should re-run the experiments with at least three different seeds, report mean plus/minus standard deviation, and verify that the reported rankings are stable across seeds.
  2. [Section IV.E (Hyperparameters tuning strategies) and Tables I-II] Hyperparameters beta, tau, lambda_2, R, and B were selected using Fig. 10, whose supporting tasks are '->Amazon' (Office-31), '->Artistic' (Office-Home), and '->Real' (DomainNet). These exact tasks appear in the evaluation tables (Table I for Office-31 and Office-Home; Table II for DomainNet). Tuning on the evaluation tasks introduces selection bias that inflates the reported accuracies for those tasks, and because the chosen hyperparameters are then applied to all seven datasets, the bias can propagate to the other tables. The tuning should be moved to held-out validation tasks or to validation splits of the target data, and the final tables should be produced without having used the reported tasks for hyperparameter selection.
  3. [Section IV.B, Section IV.E item (3), Fig. 10(a), and conclusion] The reported configuration is internally inconsistent regarding the selection ratio R. Section IV.B sets R = 0.4, and the concluding paragraph of Section IV.E repeats R = 0.4, but the joint tuning analysis in Section IV.E item (3) explicitly states that for both examined tasks 'the optimal performance was achieved with B = 32 and R = 0.3,' as shown in Fig. 10(a). This contradiction makes the exact configuration used to produce Tables I-VI ambiguous, and it undermines reproducibility. The authors must resolve the discrepancy and, ideally, re-run the final experiments under the configuration they actually claim to be optimal.
  4. [Section III.B, Eqs. (6)-(9)] The treatment of low-confidence target samples in the cluster-level losses is unspecified. Pseudo-labels are defined in Section III as onehot(argmax(P)) multiplied by the indicator 1(z > tau), so for samples with confidence below the threshold tau the pseudo-label vector is all zeros. Nevertheless, Eq. (6) assigns each target sample to a pseudo-class cluster B_{dhat}^t and includes a repulsion term -||phi_{H,k}^{s_m} - phi_{H,dhat != k}^t||^2 that pushes the source class center away from the target center of every different pseudo-class. Early in training, when target predictions are noisy and mostly below tau = 0.6, this repulsion can separate features of the same real class across domains instead of aligning them. Relatedly, the description of the Pseudo-Label Matrix in Section III.B claims that diagonal entries of PL_B^2 are always 1, yet a thresholded zero pseudo-label would yield a zero diagonal entry under the stated definition PL_B^2 = PL_B * PL_B^T. The paper should specify how below-threshold samples enter Eqs. (6)-(9) (for example, by excluding them from the cluster centers and from the PLM), and it should provide evidence that the repulsion term is not harmful in early training; the corruption experiment in Fig. 11(c) tests only the intra-domain PCM term, not the inter-domain repulsion.
minor comments (7)
  1. [Section IV.B] The lambda_1 schedule is written as '2/exp(-theta p) - 1', which is not the DANN schedule cited as reference [2]; the standard form is 2/(1+exp(-gamma p)) - 1. As printed, the expression grows without bound. Please correct the formula and define the progress variable p explicitly.
  2. [Table X] The 'L2 + Ours' row contains a corrupted entry ('62.8 53.7 4 74.4 66.1'), and the four displayed domain values are inconsistent with the stated average of 66.1. Please regenerate the table and verify all reported numbers.
  3. [Section headings] There are two sections numbered V, namely 'V. DISCUSSION' and 'V. CONCLUSION'; the conclusion should be renumbered as Section VI.
  4. [Eq. (6)] The subscript notation is inconsistent: phi_{H,i}^{s_m} is used in the definition of the cluster mean phi_{H,k}^{s_m}, while the weighted feature on the same line is defined as phi_{S,i}^{s_m}. These subscripts should be unified.
  5. [Section II.A and reference list] The sentence 'Nguyen et al. [58] proposed a cross-domain kernel classifier and applied the max-margin principle to enhance software vulnerability detection' cites [58], which is a class-aware optimal transport paper; the described work matches reference [57]. Please correct the citation.
  6. [Algorithm 1] The Input line of Algorithm 1 lists 'labeled target batch B_t'; the target batch is unlabeled in the MDA setting and should be described as such.
  7. [Section IV.E (concluding paragraph) and Section V] The statement that the chosen hyperparameters 'delivering state-of-the-art performance in each' dataset is stronger than the tables support: on DomainNet, A3MDA is below the best per-task accuracy on '->Clp' (71.4 vs. 74.3 for MCC-DA) and '->Skt' (60.7 vs. 60.8), and on Digits-5 it is below STEM on '->syn' (96.2 vs. 97.5). Additionally, Section V acknowledges that matching AHMs to scenarios required manual experimentation; the Fig. 4 ablation supports the chosen assignments, but a sentence explaining why each assignment is preferred a priori would strengthen the presentation.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the hardness feedback loop is iterative self-training, self-citations are not load-bearing, and the benchmark claim is empirical rather than definitionally forced.

full rationale

The paper's central deliverable is an empirical MDA method, not an analytic derivation from a fixed premise. The hardness values in Eqs. (1)-(3) are functions of the model's own predictions, and they are used to modulate augmentation intensity, MMD weights, and pseudo-contrastive selection in Eqs. (4)-(9). This is a standard self-training/self-supervision feedback loop: the model's outputs (and their EMA-smoothed transforms) determine the training signal, which then updates the model. There is no equation in which the reported accuracy or 'outperforms' claim is defined as an input, and no fitted scalar is renamed as a prediction; the only external evidence for the claim is the seven benchmark tables. The self-citations [75] and [77] appear in the related-work sentence 'divergence losses, particularly Maximum Mean Discrepancy (MMD) and its variants, have been widely studied [13], [72]-[77]' and are not load-bearing for any design choice or result. The Discussion's limitations (manual AHM-scenario matching and complicated hyperparameter selection) are acknowledged rather than hidden. Section IV.B does tune hyperparameters on tasks ('Amazon', 'Artistic', 'Real') that appear in Tables I and II, and the paper reports no variance and gives conflicting R values (0.3 in Fig. 10(a) vs 0.4 in the text and conclusion); these are experimental-validity and reproducibility concerns, not circularity, because the benchmark numbers are not forced by construction from the tuned values. I therefore find no significant circularity.

Assumptions & free parameters 8 free parameters · 5 assumptions · 3 invented entities

The method rests on a collection of assumptions about pseudo-label reliability, the boundedness of hardness values, the validity of a one-epoch memory delay, the manual assignment of AHMs to scenarios, and the correlation between hardness and generalization. None of these is formally proven. The free parameters are tuned on a subset of the benchmark tasks, which introduces a selection bias. The invented entities are all algorithmic constructs with no external falsifiable predictions.

free parameters (8)
  • Smoothing factor beta = 0.8
    EMA smoothing factor in Eq. (2), tuned on three tasks (Amazon from Office-31, Artistic from Office-Home, Real from DomainNet) and fixed for all datasets.
  • Pseudo-label threshold tau = 0.6
    Confidence threshold for generating pseudo-labels, tuned on the same three tasks and fixed for all datasets.
  • Intra-loss weight lambda_2 = 0.7
    Weight for L_intra in Eq. (10), tuned on three tasks and fixed for all datasets.
  • Selection ratio R = 0.4
    Fraction of hardest target samples retained for the pseudo-contrastive matrix (PCM). Note that Fig. 10(a) discussion states optimal R=0.3, but the final configuration uses R=0.4, creating an inconsistency.
  • Batch size B = 32
    Batch size, jointly tuned with R; B=32 chosen as optimal in Fig. 10(a).
  • Softmax temperature tem = 0.15
    Temperature in the PCM softmax (Eq. 8), fixed.
  • Inter-domain loss weight lambda_1 = scheduled: 2/(1+exp(-10p))-1
    Follows the DANN [2] schedule, not fitted but chosen as a standard schedule.
  • Weight eta for weighted MMD term = 1.0
    Introduced during tuning of Eq. (7), set to 1.0 to simplify the loss (effectively omitted).
assumptions (5)
  • domain assumption Pseudo-labels for all target samples are accurate enough for cluster-level alignment in Eq. (6).
    Eq. (6) uses pseudo-labels of every target sample to define same-class and different-class clusters; early in training these are noisy, and the push term may misalign true same-class features.
  • standard math Sample hardness (Basic AHM) is bounded in [0,1], making the weak/strong augmentation interpolation in Eq. (4) a valid convex combination.
    Omega_b is the L2 norm of a probability vector with one element removed; it lies in [0,1], but the paper does not state this, though it relies on it for h in Eq. (4).
  • domain assumption Hardness values from the previous epoch stored in M_H remain representative for the current epoch's augmentation decisions.
    Eq. (4) uses [S]' from memory; a one-epoch delay is assumed harmless.
  • ad hoc to paper The assignment of AHMs to scenarios (S for augmentation, H^c for inter-domain, H for intra-domain) is optimal.
    The authors admit in the Limitations that manual experimentation is required to match AHMs to scenarios; the choice is not derived from first principles.
  • domain assumption Decreasing hardness over training reflects increasing model generalization, so augmentation intensity can safely increase.
    This is the core intuition of the adaptive augmentation; it is plausible but not proven.
invented entities (3)
  • Adaptive Hardness Measurements (Basic, Smooth, Comparative AHMs)
    purpose: Quantify sample-specific hardness for guiding data augmentation, inter-domain alignment, and intra-domain alignment.
    Algorithmic constructs defined in Section III.A; they are not externally observable quantities.
  • Hardness Memory (M_H)
    purpose: Store smoothed hardness values per sample across epochs so that previous-epoch values are available for augmentation and weighting.
    A data structure described in Section III.A; no external falsifiable handle.
  • Pseudo Contrastive Matrix (PCM)
    purpose: Enforce intra-domain alignment by matching a binary pseudo-label matrix to a soft similarity matrix of predictions.
    Algorithmic construct defined in Eq. (8); used as a loss term, not a measurable entity.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Adaptive Hardness-driven Augmentation and Alignment Strategies for Multi-Source Domain Adaptations." pith.science (2026). https://pith.science/paper/BPLZVZUC

@misc{pith2026250101142,
  author       = {Pith},
  title        = {Pith review of: Adaptive Hardness-driven Augmentation and Alignment Strategies for Multi-Source Domain Adaptations},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/BPLZVZUC}},
  note         = {Machine review of arXiv:2501.01142}
}
read the original abstract

Multi-source Domain Adaptation (MDA) aims to transfer knowledge from multiple labeled source domains to an unlabeled target domain. Nevertheless, traditional methods primarily focus on achieving inter-domain alignment through sample-level constraints, such as Maximum Mean Discrepancy (MMD), neglecting three pivotal aspects: 1) the potential of data augmentation, 2) the significance of intra-domain alignment, and 3) the design of cluster-level constraints. In this paper, we introduce a novel hardness-driven strategy for MDA tasks, named "A3MDA" , which collectively considers these three aspects through Adaptive hardness quantification and utilization in both data Augmentation and domain Alignment.To achieve this, "A3MDA" progressively proposes three Adaptive Hardness Measurements (AHM), i.e., Basic, Smooth, and Comparative AHMs, each incorporating distinct mechanisms for diverse scenarios. Specifically, Basic AHM aims to gauge the instantaneous hardness for each source/target sample. Then, hardness values measured by Smooth AHM will adaptively adjust the intensity level of strong data augmentation to maintain compatibility with the model's generalization capacity.In contrast, Comparative AHM is designed to facilitate cluster-level constraints. By leveraging hardness values as sample-specific weights, the traditional MMD is enhanced into a weighted-clustered variant, strengthening the robustness and precision of inter-domain alignment. As for the often-neglected intra-domain alignment, we adaptively construct a pseudo-contrastive matrix by selecting harder samples based on the hardness rankings, enhancing the quality of pseudo-labels, and shaping a well-clustered target feature space. Experiments on multiple MDA benchmarks show that " A3MDA " outperforms other methods.

Figures

Figures reproduced from arXiv: 2501.01142 by the authors.

Figure 1
Figure 1. Red boxes indicate aspects that are not handled effectively, while green boxes signify effectively handled aspects. The core of A3 MDA lies in using three-fold Adaptive hardness quantification and utilization to address three limitations [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. (a) Diagram of our A3 MDA framework. Consistent with [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

79 extracted references · 74 canonical work pages

  1. [1]

    Deep coral: Correlation alignment for deep domain adaptation,

    B. Sun and K. Saenko , “Deep coral: Correlation alignment for deep domain adaptation,” in Computer Vision –ECCV 2016 Workshops: Amsterdam, The Netherlands, October 8 -10 and 15 -16, 2016, Proceedings, Part III 14. Springer, 2016, pp. 443–450

    work page 2016

  2. [2]

    Unsupervised domain adaptation by backpropagation,

    Y. Ganin and L. Victor, “Unsupervised domain adaptation by backpropagation,” International conference on machine learning . PMLR, 2015

    work page 2015

  3. [3]

    Deep subdomain adaptation network for image classification,

    Y. Zhu, F. Zhuang, J. Wang, G. Ke, J. Chen, J. Bian, H. Xiong, and Q. He, “Deep subdomain adaptation network for image classification,” IEEE transactions on neural networks and learning systems, vol. 32, no. 4, pp. 1713–1722, 2020

    work page 2020

  4. [4]

    Information -theoretic regularization for multi-source domain adaptation,

    G. Y. Park and S. W. Lee, “Information -theoretic regularization for multi-source domain adaptation,” in Proceedings of the IEEE/CVF International Conference on Computer Vision, 2021, pp. 9214–9223

    work page 2021

  5. [5]

    Moment matching for multi- source domain adaptation,

    X. Peng, Q. Bai, X. Xia, Z. Huang, K. Saenko, and B. Wang, “Moment matching for multi- source domain adaptation,” in Proceedings of the IEEE/CVF international conference on computer vision , 2019, pp. 1406–1415

    work page 2019

  6. [6]

    Multi -source distilling domain adaptation,

    S. Zhao, G. Wang, S. Zhang, Y. Gu, Y. Li, Z. Song, P. Xu, R. Hu, H. Chai, and K. Keutzer, “Multi -source distilling domain adaptation,” in Proceedings of the AAAI Conference on Artificial Intelligence , vol. 34, no. 07, 2020, pp. 12 975–12 983

    work page 2020

  7. [7]

    Multi -source domain adaptation with mixture of joint distributions,

    S. Chen, “Multi -source domain adaptation with mixture of joint distributions,” Pattern Recognition, p. 110295, 2024

    work page 2024

  8. [8]

    Connecting the dots with landmarks: Discriminatively learning domain -invariant features for unsupervised domain adaptation,

    B. Gong, K. Grauman, and F. Sha, “Connecting the dots with landmarks: Discriminatively learning domain -invariant features for unsupervised domain adaptation,” in International conference on machine learning . PMLR, 2013, pp. 222–230

    work page 2013

Show all 79 references
  1. [9]

    Dynamic instance domain adaptation,

    Z. Deng, K. Zhou, D. Li, J. He, Y.- Z. Song, and T. Xiang, “Dynamic instance domain adaptation,” IEEE Transactions on Image Processing , vol. 31, pp. 4585–4597, 2022

    2022

  2. [10]

    Prototype- based multisource domain adaptation,

    L. Zhou, M. Ye, D. Zhang, C. Zhu, and L. Ji, “Prototype- based multisource domain adaptation,” IEEE Transactions on Neural Networks and Learning Systems, vol. 33, no. 10, pp. 5308–5320, 2021

    2021

  3. [11]

    Deep cocktail network: Multi-source unsupervised domain adaptation with category shift,

    R. Xu, Z. Chen, W. Zuo, J. Yan, and L. Lin, “Deep cocktail network: Multi-source unsupervised domain adaptation with category shift,” in Proceedings of the IEEE conference on computer vision and pattern recognition, 2018, pp. 3964–3973

    2018

  4. [12]

    Correlation alignment for unsupervised domain adaptation,

    B. Sun, J. Feng, and K. Saenko, “Correlation alignment for unsupervised domain adaptation,” Domain adaptation in computer vision applications, pp. 153–171, 2017

    2017

  5. [13]

    Learning transferable features with deep adaptation networks,

    M. Long, Y. Cao, J. Wang, and M. Jordan, “Learning transferable features with deep adaptation networks,” in International conference on machine learning. PMLR, 2015, pp. 97–105

    2015

  6. [14]

    Deep transfer learning with joint adaptation networks,

    M. Long, H. Zhu, J. Wang, and M. I. Jordan, “Deep transfer learning with joint adaptation networks,” in International conference on machine learning. PMLR, 2017, pp. 2208–2217

    2017

  7. [15]

    Mutual learning of joint and separate domain alignments for multi -source domain adaptation,

    Y. Xu, M. Kan, S. Shan, and X. Chen, “Mutual learning of joint and separate domain alignments for multi -source domain adaptation,” in Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, 2022, pp. 1890–1899

    2022

  8. [16]

    A simple baseline for semi-supervised semantic segmentation with strong data augmentation,

    J. Yuan, Y. Liu, C. Shen, Z. Wang, and H. Li, “A simple baseline for semi-supervised semantic segmentation with strong data augmentation,” in Proceedings of the IEEE/CVF International Conference on Computer Vision, 2021, pp. 8229–8238

    2021

  9. [17]

    Multi- representation adaptation network for cross-domain image classification,

    Y. Zhu, F. Zhuang, J. Wang, J. Chen, Z. Shi, W. Wu, and Q. He, “Multi- representation adaptation network for cross-domain image classification,” Neural Networks, vol. 119, pp. 214–221, 2019

    2019

  10. [18]

    Aligning domain -specific distribution and classifier for cross -domain classification from multiple sources,

    Y. Zhu, F. Zhuang, and D. Wang, “Aligning domain -specific distribution and classifier for cross -domain classification from multiple sources,” in AAAI conference on artificial intelligence , vol. 33, no. 01, 2019, pp. 5989–5996

    2019

  11. [19]

    Mind the class weight bias: Weighted maximum mean discrepancy for unsupervised domain adaptation,

    H. Yan, Y. Ding, P. Li, Q. Wang, Y. Xu, and W. Zuo, “Mind the class weight bias: Weighted maximum mean discrepancy for unsupervised domain adaptation,” in Proceedings of the IEEE conference on computer vision and pattern recognition, 2017, pp. 2272–2281

    2017

  12. [20]

    Advent: Adversarial entropy minimization for domain adaptation in semantic segmentation,

    T.-H. Vu, H. Jain, M. Bucher, M. Cord, and P. P´erez, “Advent: Adversarial entropy minimization for domain adaptation in semantic segmentation,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2019, pp. 2517–2526

    2019

  13. [21]

    Adversarial entropy optimization for unsupervised domain adaptation,

    A. Ma, J. Li, K. Lu, L. Zhu, and H. T. Shen, “Adversarial entropy optimization for unsupervised domain adaptation,” IEEE Transactions on Neural Networks and Learning Systems , vol. 33, no. 11, pp. 6263 – 6274, 2021

    2021

  14. [22]

    Active adversarial domain adaptation,

    J.-C. Su, Y. -H. Tsai, K. Sohn, et al. , “Active adversarial domain adaptation,” in Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, 2020, pp. 739–748

    2020

  15. [23]

    Domain adaptation via incremental confidence samples into classification,

    S. Teng, Z. Zheng, N. Wu, L. Fei, and W. Zhang, “Domain adaptation via incremental confidence samples into classification,” International Journal of Intelligent Systems, vol. 37, no. 1, pp. 365–385, 2022

    2022

  16. [24]

    Pseudo -loss confidence metric for semi -supervised few -shot learning,

    K. Huang, J. Geng, W. Jiang, X. Deng, and Z. Xu, “Pseudo -loss confidence metric for semi -supervised few -shot learning,” in Proceedings of the IEEE/CVF International Conference on Computer Vision, 2021, pp. 8671–8680

    2021

  17. [25]

    Adversarial discriminative domain adaptation,

    E. Tzeng, J. Hoffman, K. Saenko, et al. , “Adversarial discriminative domain adaptation,” in Proceedings of the IEEE conference on computer vision and pattern recognition, 2017, pp. 7167–7176

    2017

  18. [26]

    Coupled generative adversarial networks,

    M.-Y. Liu and O. Tuzel, “Coupled generative adversarial networks,” Advances in neural information processing systems, vol. 29, 2016

    2016

  19. [27]

    Maximum classifier discrepancy for unsupervised domain adaptation,

    K. Saito, K. Watanabe, Y. Ushiku, and T. Harada, “Maximum classifier discrepancy for unsupervised domain adaptation,” in Proceedings of the IEEE conference on computer vision and pattern recognition , 2018, pp. 3723–3732

    2018

  20. [28]

    Multi -source contribution learning for domain adaptation,

    K. Li, J. Lu, H. Zuo, and G. Zhang, “Multi -source contribution learning for domain adaptation,” IEEE Transactions on Neural Networks and Learning Systems, vol. 33, no. 10, pp. 5293–5307, 2021

    2021

  21. [29]

    Stem: An approach to multi -source domain adaptation with guarantees,

    V.-A. Nguyen, T. Nguyen, T. Le, Q. H. Tran, and D. Phung, “Stem: An approach to multi -source domain adaptation with guarantees,” in Proceedings of the IEEE/CVF International Conference on Computer Vision, 2021, pp. 9352–9363

    2021

  22. [30]

    Hard -aware deeply cascaded embedding,

    Y. Yuan, K. Yang, and C. Zhang, “Hard -aware deeply cascaded embedding,” in Proceedings of the IEEE international conference on computer vision, 2017, pp. 814–823

    2017

  23. [31]

    Model adaptation with synthetic and real data for semantic dense foggy scene understanding,

    C. Sakaridis, D. Dai, S. Hecker, and L. Van Gool, “Model adaptation with synthetic and real data for semantic dense foggy scene understanding,” in Proceedings of the european conference on computer vision (ECCV), 2018, pp. 687–704

    2018

  24. [32]

    Unsupervised intra-domain adaptation for semantic segmentation through self - supervision,

    F. Pan, I. Shin, F. Rameau, S. Lee, and I. S. Kweon, “Unsupervised intra-domain adaptation for semantic segmentation through self - supervision,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2020, pp. 3764–3773

    2020

  25. [33]

    Instance credibility inference for few - shot learning,

    Y. Wang, C. Xu, C. Liu, et al., “Instance credibility inference for few - shot learning,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2020, pp. 12 836–12 845

    2020

  26. [34]

    Instance -specific and model-adaptive supervision for semi -supervised semantic segmentation,

    Z. Zhao, S. Long, J. Pi, J. Wang, and L. Zhou, “Instance -specific and model-adaptive supervision for semi -supervised semantic segmentation,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2023, pp. 23 705–23 714

    2023

  27. [35]

    Towards fewer annotations: Active learning via region impurity and prediction uncertainty for domain adaptive semantic segmentation,

    B. Xie, L. Yuan, S. Li, C. H. Liu, and X. Cheng, “Towards fewer annotations: Active learning via region impurity and prediction uncertainty for domain adaptive semantic segmentation,” in Proceedings 15 of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2022...

    2022

  28. [36]

    Mixstyle neural networks for domain generalization and adaptation,

    K. Zhou, Y. Yang, Y. Qiao, and T. Xiang, “Mixstyle neural networks for domain generalization and adaptation,” International Journal of Computer Vision, pp. 1–15, 2023

    2023

  29. [37]

    mixup: Beyond empirical risk minimization,

    H. Zhang, M. Cisse, Y. N. Dauphin, and D. Lopez-Paz, “mixup: Beyond empirical risk minimization,” in International Conference on Learning Representations, 2018. [Online]. Available: https://openreview.net/forum?id=r1Ddp1-Rb

    2018

  30. [38]

    Cutmix: Regularization strategy to train strong classifiers with localizable features,

    S. Yun, D. Han, S. J. Oh, S. Chun, J. Choe, and Y. Yoo, “Cutmix: Regularization strategy to train strong classifiers with localizable features,” in Proceedings of the IEEE/CVF international conference on computer vision, 2019, pp. 6023–6032

    2019

  31. [39]

    Autoaugment: Learning augmentation strategies from data,

    E. D. Cubuk, B. Zoph, D. Mane, V. Vasudevan, and Q. V. Le, “Autoaugment: Learning augmentation strategies from data,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2019, pp. 113–123

    2019

  32. [40]

    Randaugment: Practical automated data augmentation with a reduced search space,

    E. D. Cubuk, B. Zoph, J. Shlens, and Q. V. Le, “Randaugment: Practical automated data augmentation with a reduced search space,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition workshops, 2020, pp. 702–703

    2020

  33. [41]

    Multi -source unsupervised domain adaptation via pseudo target domain,

    C.-X. Ren, Y. -H. Liu, X. -W. Zhang, and K. -K. Huang, “Multi -source unsupervised domain adaptation via pseudo target domain,” IEEE Transactions on Image Processing, vol. 31, pp. 2122–2135, 2022

    2022

  34. [42]

    Learning to combine: Knowledge aggregation for multi -source domain adaptation,

    H. Wang, M. Xu, B. Ni, et al., “Learning to combine: Knowledge aggregation for multi -source domain adaptation,” in Computer Vision– ECCV 2020: 16th European Conference, Glasgow, UK, August 23 –28, 2020, Proceedings, Part VIII 16. Springer, 2020, pp. 727–744

    2020

  35. [43]

    A new progressive multisource domain adaptation network with weighted decision fusion,

    Z.-G. Liu, L. -B. Ning, and Z. -W. Zhang, “A new progressive multisource domain adaptation network with weighted decision fusion,” IEEE Transactions on Neural Networks and Learning Systems, 2022

    2022

  36. [44]

    T-svdnet: Exploring high-order prototypical correlations for multi -source domain adaptation,

    R. Li, X. Jia, J. He, S. Chen, and Q. Hu, “T-svdnet: Exploring high-order prototypical correlations for multi -source domain adaptation,” in Proceedings of the IEEE/CVF International Conference on Computer Vision, 2021, pp. 9991–10 000

    2021

  37. [45]

    Self -paced supervision for multi- source domain adaptation,

    Z. Wang, C. Zhou, B. Du, and F. He, “Self -paced supervision for multi- source domain adaptation,” in Proceedings of the Thirty -First International Joint Conference on Artificial Intelligence, IJCAI-22, 2022

    2022

  38. [46]

    Domain -specific feature elimination: multi- source domain adaptation for image classification,

    K. Wu, F. Jia, and Y. Han, “Domain -specific feature elimination: multi- source domain adaptation for image classification,” Frontiers of Computer Science, vol. 17, no. 4, p. 174705, 2023

    2023

  39. [47]

    Adapting visual category models to new domains,

    K. Saenko, B. Kulis, M. Fritz, et al., “Adapting visual category models to new domains,” in Computer Vision –ECCV 2010: 11th European Conference on Computer Vision, Heraklion, Crete, Greece, September 5-11, 2010, Proceedings, Part IV 11. Springer, 2010, pp. 213–226

    2010

  40. [48]

    Deep hashing network for unsupervised domain adaptation,

    H. Venkateswara, J. Eusebio, S. Chakraborty, and S. Panchanathan, “Deep hashing network for unsupervised domain adaptation,” in Proceedings of the IEEE conference on computer vision and pattern recognition, 2017, pp. 5018–5027

    2017

  41. [49]

    Automix: Unveiling the power of mixup for stronger classifiers,

    Z. Liu, S. Li, D. Wu, Z. Liu, Z. Chen, L. Wu, and S. Z. Li, “Automix: Unveiling the power of mixup for stronger classifiers,” in ECCV 2022 . Springer, 2022, pp. 441–458

    2022

  42. [50]

    Training multi -source domain adaptation network by mutual information estimation and minimization,

    L. Wen, S. Chen, M. Xie, C. Liu, and L. Zheng, “Training multi -source domain adaptation network by mutual information estimation and minimization,” Neural Networks, vol. 171, pp. 353–361, 2024

    2024

  43. [51]

    Semi-sup

    is a dataset that includes images from four domains: Art (A), Cartoon (C), Sketch (S), and Photo (P). Each domain contains 7 categories, with the following number of images: P (1,670 images), A (2,048 images), C (2,048 images), and S (3,929 images). Digits-5 [5] dataset consis...

  44. [52]

    Deeper, broader and artier domain generalization,

    D. Li, Y. Yang, Y. -Z. Song, and T. M. Hospedales, “Deeper, broader and artier domain generalization,” in Proceedings of the IEEE international conference on computer vision, 2017, pp. 5542–5550

    2017

  45. [53]

    Geodesic flow kernel for unsupervised domain adaptation,

    B. Gong, Y. Shi, F. Sha, and K. Grauman, “Geodesic flow kernel for unsupervised domain adaptation,” in 2012 IEEE conference on computer vision and pattern recognition. IEEE, 2012, pp. 2066–2073

    2012

  46. [54]

    Deep residual learning for image recognition,

    K. He, X. Zhang, S. Ren, and J. Sun, “Deep residual learning for image recognition,” in Proceedings of the IEEE conference on computer vision and pattern recognition, 2016, pp. 770–778

    2016

  47. [55]

    Weighted correlation embedding learning for domain adaptation,

    Lu Y, Zhu Q, Zhang B, et al., “Weighted correlation embedding learning for domain adaptation,” IEEE Transactions on Image Processing, vol.31, pp. 5303-5316, 2022

    2022

  48. [56]

    Low-Rank Correlation Learning for Unsupervised Domain Adaptation ,

    Lu Y, Wong W K, Yuan C, et al. , “Low-Rank Correlation Learning for Unsupervised Domain Adaptation ,” IEEE Transactions on Multimedia , 2023

    2023

  49. [57]

    Guided discrimination and correlation subspace learning for domain adaptation ,

    Lu Y, Wong W K, Zeng B, et al., “Guided discrimination and correlation subspace learning for domain adaptation ,” IEEE Transactions on Image Processing, vol.32, pp. 2017-2032, 2023

    2017

  50. [58]

    Deep Domain Adaptation With Max -Margin Principle for Cross -Project Imbalanced Software Vulnerability Detection ,

    Nguyen V, Le T, Tantithamthavorn C, et al. , “Deep Domain Adaptation With Max -Margin Principle for Cross -Project Imbalanced Software Vulnerability Detection ,” ACM Transactions on Software Engineering and Methodology, 2024

    2024

  51. [59]

    A Class-aware Optimal Transport Approach with Higher -Order Moment Matching for Unsupervised Domain Adaptation,

    Nguyen T, Nguyen V, Le T, et al. , “A Class-aware Optimal Transport Approach with Higher -Order Moment Matching for Unsupervised Domain Adaptation,” arXiv preprint arXiv:2401.15952, 2024

    2024 arXiv

  52. [60]

    Test-time Adaptation against Multi- modal Reliability Bias ,

    Yang M, Li Y, Zhang C, et al., “Test-time Adaptation against Multi- modal Reliability Bias ,” The Twelfth International Conference on Learning Representations. 2024

    2024

  53. [61]

    Cross-domain adaptive clustering for semi - supervised domain adaptation ,

    Li J, Li G, Shi Y, et al. , “Cross-domain adaptive clustering for semi - supervised domain adaptation ,” Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2021, pp. 2505-2514

    2021

  54. [62]

    A discriminatively deep fusion approach with improved conditional GAN (im -cGAN) for facial expression recognition,

    Sun Z, Zhang H, Bai J, et al. , “A discriminatively deep fusion approach with improved conditional GAN (im -cGAN) for facial expression recognition,” Pattern Recognition, 135, pp. 109157, 2023

    2023

  55. [63]

    Robust object re -identification with coupled noisy labels,

    Yang M, Huang Z , et al., “Robust object re -identification with coupled noisy labels,” International Journal of Computer Vision, pp. 1-19, 2024

    2024

  56. [64]

    Graph matching with bi -level noisy correspondence,

    Lin Y, Yang M, Yu J, et al. , “Graph matching with bi -level noisy correspondence,” in Proceedings of the IEEE/CVF international conference on computer vision, 2023, pp. 23362-23371

    2023

  57. [65]

    Robust multi -view clustering with incomplete information ,

    Yang M, Li Y, Hu P, et al. , “Robust multi -view clustering with incomplete information ,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 45, no. 1, pp. 1055-1069, 2022

    2022

  58. [66]

    Rethinking the inception architecture for computer vision,

    C. Szegedy, V. Vanhoucke, S. Ioffe, J. Shlens, and Z. Wojna, “Rethinking the inception architecture for computer vision, ” in Proceedings of the IEEE Conference on computer vision and pattern recognition, 2016, pp. 2818-2826

    2016

  59. [67]

    Densely connected convolutional networks,

    G. Huang, Z. Liu, L. Van Der Maaten, et al., “Densely connected convolutional networks, ” in Proceedings of the IEEE conference on computer vision and pattern recognition, 2017, pp. 4700-4708

    2017

  60. [68]

    An image is worth 16x16 words: Transformers for image recognition at scale,

    A. Dosovitskiy, L. Beyer, A. Kolesnikov , et al. , “An image is worth 16x16 words: Transformers for image recognition at scale,” in International Conference on Learning Representations, 2021

    2021

  61. [69]

    Euclidean distance mapping ,

    Danielsson, P. E., “Euclidean distance mapping ,” Computer Graphics and image processing, 14(3), 227-248, 1980

    1980

  62. [70]

    Wasserstein distance guided representation learning for domain adaptation,

    J. Shen, Y. Qu, W. Zhang, and Y. Yu, “ Wasserstein distance guided representation learning for domain adaptation, ” in Proceedings of the AAAI conference on artificial intelligence, 2018, vol. 32, no. 1

    2018

  63. [71]

    Federated Domain Adaptation via Transformer for Multi -site Alzheimer’s Disease Diagnosis,

    Lei B, Zhu Y, Liang E, et al., “ Federated Domain Adaptation via Transformer for Multi -site Alzheimer’s Disease Diagnosis, ” IEEE Transactions on Medical Imaging, 2023

    2023

  64. [72]

    Graph Convolution and Self -attention Enhanced CNN with Domain Adaptation for Multi -site COVID- 19 Diagnosis,

    Zhang J, Liu Y, Lei B, et al., “Graph Convolution and Self -attention Enhanced CNN with Domain Adaptation for Multi -site COVID- 19 Diagnosis,” 45th Annual International Conference of the IEEE Engineering in Medicine & Biology Society, 2023: 1-4

    2023

  65. [73]

    A new progressive multisource domain adaptation network with weighted decision fusion,

    Z. Liu, L. -B. Ning, and Z. -W. Zhang, “A new progressive multisource domain adaptation network with weighted decision fusion,” IEEE Trans. Neural Netw. Learn. Syst., vol. 35, no. 1, pp. 1062–1072, 2022

    2022

  66. [74]

    Multi -source collaborative contrastive learning for decentralized domain adaptation,

    Y. Wei, L. Yang, Y. Han, and Q. Hu, “Multi -source collaborative contrastive learning for decentralized domain adaptation,” IEEE Trans. Circuits Syst. Video Technol., vol. 33, no. 5, pp. 2202–2216, 2022

    2022

  67. [75]

    Multi -source transfer learning via optimal transport feature ranking for EEG classification,

    J. Li, Q. She, F. Fang, Y. Chen, and Y. Zhang, “Multi -source transfer learning via optimal transport feature ranking for EEG classification,” Neurocomputing, vol. 596, p. 127944, 2024, doi: 10.1016/j.neucom.2024.127944

    2024

  68. [76]

    DANE: A dual -level alignment network with ensemble learning for multi -source domain adaptation,

    Y. Yang, L. Wen, P. Zeng, B. Yan, and Y. Wang, “DANE: A dual -level alignment network with ensemble learning for multi -source domain adaptation,” IEEE Trans. Instrum. Meas., vol. 73, pp. 2456-2467, 2024

    2024

  69. [77]

    Adaptive intermediate class -wise distribution alignment: A universal domain adaptation and generalization method for machine fault diagnosis,

    Q. Qian, J. Luo, and Y. Qin, “Adaptive intermediate class -wise distribution alignment: A universal domain adaptation and generalization method for machine fault diagnosis,” IEEE Trans. Neural Networks Learn. Syst., vol. 35, no. 5, pp. 1234-1247, May 2024

    2024

  70. [78]

    Learning with alignments: Tackling the inter - and intra-domain shifts for cross -multidomain facial expression recognition,

    Y. Yang, L. Wen, X. Zeng, Y. Xu, X. Wu, J. Zhou, and Y. Wang, “Learning with alignments: Tackling the inter - and intra-domain shifts for cross -multidomain facial expression recognition,” in Proc. 32nd ACM Int. Conf. Multimedia, Oct. 2024, pp. 4236-4245

    2024

  71. [79]

    Domain adaptation in reinforcement learning: a comprehensive and systematic study,

    A. Farhadi, M. Mirzarezaee, A. Sharifi, et al., “Domain adaptation in reinforcement learning: a comprehensive and systematic study,” Front. Inf. Technol. Electron. Eng., vol. 25, no. 11, pp. 1446–1465, 2024

    2024

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.