Pith. sign in

REVIEW 4 major objections 5 minor 39 references

Locating and Mitigating Gradient Conflicts in Point Cloud Domain Adaptation via Saliency Map Skewness

T0 review · 4 major / 5 minor · reviewed 2026-08-16 · deepseek-v4-flash

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

Pith's one-line read Saliency-map skewness can identify the self-supervision gradients that damage point-cloud domain adaptation, and a plug-in block that filters those samples out raises classification accuracy and segmentation transfer.

desk verdict A plausible plug-in for point cloud UDA whose causal mechanism (skewness = gradient conflict) is not established; worth peer review but needs major revisions. read the letter →

arxiv 2504.15796 v1 pith:6ACVA4HI submitted 2025-04-22 cs.CV cs.LG

Jiaqi Tang , Yinsong Xu , Qingchao Chen This is my paper · ORCID
classification cs.CVcs.LG
keywords pointcloudunsuperviseddomainadaptationgradientconflictsaliencymapskewnessself-supervisionsampleselectionmulti-tasklearning
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

The paper sets out to show that in multi-task point cloud unsupervised domain adaptation (UDA), the self-supervision gradients that usually help alignment are harmful for some samples, and that this harm can be detected without target labels. Its proposal, the Saliency Map-based Data Sampling Block (SM-DSB), estimates per-sample gradient conflict from the skewness of the 3D saliency map and drops high-skewness samples from the self-supervision loss. If the claim holds, any point cloud UDA method built on a classification-plus-self-supervision framework can gain accuracy and robustness by adding a lightweight plug-in instead of redesigning the task. The paper reports that plugging SM-DSB into Self-dist GCN raises the PointDA-10 average accuracy from 75.1% to 77.2%, and that it also improves segmentation transfer on PointSegDA.

What carries the argument

The load-bearing object is the skewness of the point cloud saliency map. The saliency map assigns each point $p_i$ a score $s_i = -\partial L / \partial r_i \, r_i^{1+\alpha}$, measuring how much the classification loss would rise if the point moved toward the cloud's core in spherical coordinates; skewness is then the standardized third moment $sk = \sum_i (S_i - \bar S)^3 / (n \sigma^3)$ of these scores. The SM-DSB has two parts: a Measurer that computes $sk$ for each sample in a batch through an extra gradient back-propagation, and a Selector that sorts the scores, sets a threshold in the batch's skewness distribution, and assigns weight $\lambda_b = 0$ to self-supervision loss terms for samples at or above the threshold, excluding the high-skewness tail. The paper uses this selection in the first training stage on source data and in the second stage on target pseudo-labels, adding a small Gaussian perturbation to pseudo-label skewness scores to offset the noise the paper measures in those labels.

What would settle it

On a fixed training run, compute for every target sample the cosine similarity between its self-supervision gradient and the oracle gradient from true target labels, then bin samples by saliency-map skewness. If the highest-skewness bin does not show systematically lower or negative cosine similarity than the lowest-skewness bin across the whole training trajectory, skewness is not a reliable conflict estimator and the selection rule's stated mechanism would fail.

Watch

Extended reading notes

Core claim

The central claim is that gradient conflict between the auxiliary self-supervision task and the target classification objective is not uniform across samples, and that the skewness of a sample's saliency map is a reliable, label-free estimate of that conflict. A high-skewness saliency map means the model leans on a few critical points, a shortcut-like reliance that is fragile under domain shift; such samples' self-supervision gradients tend to point away from the oracle target-label gradient. SM-DSB therefore computes the skewness of each sample's saliency map, sorts the batch, and zeros out the self-supervision loss contribution for samples whose skewness exceeds an adaptive threshold, keeping only the gradients that move the summed gradient closer to the oracle. The paper argues that this both explains the inconsistent per-class gains seen with self-supervision and provides a general mitigation strategy. Evidence includes a measured correlation between skewness and the cosine similarity of self-supervision and oracle gradients, an additive-noise-model test for the causal direction, and accuracy gains on six PointDA-10 transfer tasks.

Load-bearing premise

The load-bearing premise is that saliency-map skewness reliably estimates the cosine similarity between a sample's self-supervision gradient and the oracle classification gradient, so that high skewness marks samples whose self-supervision gradients hurt the target objective.

Editorial extensions

If this is right

  • Plugged into Self-dist GCN, the block raises the average classification accuracy on PointDA-10 from 75.1% to 77.2%, with the largest gains on the hardest synthetic-to-real transfers.
  • On PointSegDA, applying the block to DefRec+PCM improves test mean IoU and wins 10 of the 12 transfer directions.
  • The block also helps when moved to a PointNet backbone and to other self-supervised UDA methods such as GAST and DefRec+PCM, supporting the claim that it is a model-agnostic plugin.
  • The analysis finds that inter-task gradient conflicts stay stable while the aggregated-gradient-versus-oracle conflict decays during training, implying that conventional multi-task gradient surgery alone cannot prevent negative transfer in UDA.
  • Because the selection only changes which samples contribute to the self-supervision loss, the method adds modest computation and can be integrated into existing frameworks without changing the task design.

Reading between the lines

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

  • If skewness really is the causal driver, the same selection rule should transfer to 2D image UDA by replacing the 3D saliency map with the standard input-gradient saliency map; a positive result there would confirm the mechanism is about gradient geometry, not point-cloud specifics.
  • The paper's own observation that the optimal selection ratio $\beta$ interacts with the training stage suggests an adaptive schedule for $\beta$ as a natural extension, possibly yielding further gains without oracle access.
  • The high-skewness samples the block drops are predicted to be exactly those that fail under geometric corruption; a stress test that corrupts the selected-away samples should show larger accuracy drops than corrupting the retained ones.
  • A competing explanation is that the block works by discarding pseudo-label noise in the target stage; comparing SM-DSB against a confidence-based filter would isolate whether skewness adds information beyond label noise.
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

4 major / 5 minor

Summary. The paper proposes SM-DSB, a plug-in module for point-cloud unsupervised domain adaptation (UDA) that uses the skewness of per-instance 3D saliency maps to estimate whether a sample's self-supervision gradient conflicts with the oracle classification gradient. Samples with high skewness are gated out of the self-supervision loss, based on the claim that high skewness correlates with large gradient conflict. The module is evaluated when plugged into Self-dist GCN, GAST, and DefRec+PCM on PointDA-10 classification and PointSegDA segmentation, reporting average accuracy/IoU improvements over the respective baselines and state-of-the-art results on several transfer directions.

Significance. If the proposed mechanism is correct, the paper offers a lightweight, model-agnostic way to reduce negative transfer in multi-task point-cloud UDA, and the gradient-conflict analysis in Section V.B is a useful conceptual contribution. The empirical improvement is plausible and consistent across several backbones, methods, and datasets, and the paper includes ablations and analyses beyond a single benchmark. However, the central mechanistic claim -- that saliency-map skewness is a per-sample proxy for gradient conflict -- is not directly verified, and the paper contains internal inconsistencies between the stated selection rule and the algorithm. The contribution is therefore promising but not yet fully supported.

major comments (4)
  1. [Section III.C, Figure 1, Section V.C] The selection rule in Eq. (6)-(7) and Algorithm 1 is applied at the per-sample level, but the evidence for the skewness-conflict correlation is only shown at the level of training steps in Figure 1, with one point per step. The ANM analysis in Section V.C reports scores of 0.16 versus 0.25 but does not state the unit of analysis, the number of observations, or any calibration of the scores. To support the load-bearing claim, the authors should directly measure, per sample within a batch, the cosine similarity between that sample's self-supervision gradient and the oracle classification gradient (computed with target labels in an analysis-only setting) and report the correlation with per-sample skewness, e.g., as a binned scatterplot or correlation coefficient with confidence intervals.
  2. [Section III.C, Section V.C] There is an internal inconsistency in the direction of the selection rule. Equation (7) and Algorithm 1 set the gating weight to zero when sk_b >= tau, meaning low-skewness samples participate in self-supervision, while the text in Section III.C says that "only those with high scores will participate in further processing." Section V.C similarly states that the model should be updated through the gradient from the more difficult task, which would imply keeping high-skewness samples. The authors should unambiguously state which samples are kept and reconcile the prose with Eq. (7), because the mechanistic interpretation depends on this direction.
  3. [Table II] The average column in Table II is internally inconsistent. For the "DefRec+ours" row, the 12 per-direction values (81.3, 61.2, 64.3, 67.8, 36.3, 71.6, 42.1, 30.0, 31.4, 73.2, 72.2, 67.2) average to 58.2, not 67.2; the "DefRec" row also does not match its reported average of 58.1 (computed: 57.1). The claimed improvement in average IoU should be corrected, and the per-direction numbers should be checked.
  4. [Table I, Section IV.C] The headline comparison in Table I is weakened by checkpoint selection on the test set, as acknowledged in the table footnote and in Section IV.C. For a fair assessment of the claimed state-of-the-art performance, the authors should report results obtained with a fixed training schedule or a validation-based stopping criterion, and should state whether the same test-set checkpoint selection was applied to the baseline methods being compared.
minor comments (5)
  1. [Section IV.A] The sentence "In this article, all models are trained on one NVIDIA 3090 GPU" is not an implementation detail relevant to reproducibility; the authors should report the optimizer, learning rate, and number of training steps per method instead.
  2. [Section V.D] The pseudo-label skewness perturbation is set to a Gaussian with mean 0.1 because the measured discrepancy in Figure 7 is 0.1, but this is a fitted correction. The authors should clarify whether this value is fixed for all datasets and why a mean shift, rather than a variance adjustment, is the appropriate model of pseudo-label noise.
  3. [Section III.C, Eq. (5)] Equation (5) uses the notation S with an overline but defines it as "the mean value of its saliency map"; the bar symbol is missing in the displayed equation. Also, the definition of sigma as the standard deviation of the saliency scores should be stated explicitly.
  4. [Figure 1] The phrase "more gain in the cousin similarity" appears to be a typo for "cosine similarity." Please check all figure captions and text for similar errors.
  5. [Throughout] The paper does not report error bars or multiple seeds for the main comparison tables. Given the small accuracy differences reported in some settings (e.g., 0.3-0.9 points), providing variance estimates or significance tests would strengthen the empirical claims.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: SM-DSB is an empirical sample-selection heuristic whose effect is measured by direct comparisons, not derived from its own inputs.

full rationale

I walked the derivation chain from the saliency-map score (Eq. 4) and skewness (Eq. 5) to the gating rule (Eq. 6-7, Algorithm 1). No equation-level reduction occurs: skewness is not defined in terms of the oracle-gradient cosine similarity, nor is the cosine similarity computed from skewness. The claim that high skewness tracks gradient conflict is supported only by correlational evidence (Fig. 1, ANM in Sec. V.C); that evidence is weak and the analysis is at training-step level rather than per sample, but weakness of evidence is not circularity. The pseudo-label perturbation mean of 0.1 (Sec. V.D) and the selection ratio beta (Sec. V.E) are calibrated hyperparameters, not quantities renamed as predictions, and the final accuracy gains are measured outcomes rather than consequences forced by these settings. There is no load-bearing self-citation: the saliency-map construction is cited to prior work [29] and is not equivalent to the gradient-conflict estimate the paper introduces. The central claim therefore has independent empirical content, and no circular step can be exhibited. I also note an internal inconsistency between the prose description and Eq. 7/Algorithm 1 about which samples are kept, but that is a consistency issue, not circularity.

Assumptions & free parameters 4 free parameters · 4 assumptions · 0 invented entities

The paper's central claim rests on an empirical correlation between saliency-map skewness and gradient conflict, plus several tuned constants (beta, perturbation mean, alpha, topIndex mask). Most of the work is a heuristic validated by experiments rather than a derivation, so the axiom ledger is mostly domain assumptions about the validity of the proxy and the pseudo-labels.

free parameters (4)
  • selection ratio beta = 0.7 for DGCNN, 0.8 for PointNet
    Controls the adaptive threshold tau; tuned per backbone/setup; ablations in Fig. 8-10 show sensitivity across transfer directions, so choosing a single value is a fitted choice.
  • pseudo-label skewness perturbation mean = 0.1 (std not reported)
    Added to target-domain skewness scores to compensate for pseudo-label noise; the value equals the observed discrepancy in Fig. 7, so it is fit to data.
  • alpha scaling factor = 1
    Scaling factor in the saliency map formula, inherited from [29]; ablation shows almost no effect, so it is a parameter but not an influential one.
  • topIndex mask percentage = 0.7
    Mentioned in implementation details with a promise of later discussion that never appears; effect on results unknown.
assumptions (4)
  • ad hoc to paper Skewness of the saliency map is a domain-agnostic proxy for the cosine similarity between self-supervision gradients and oracle classification gradients.
    Invoked in Section III.C and Fig. 1 to justify the selection rule; correlation is shown for one setup and never given a quantitative coefficient in the text.
  • domain assumption Self-supervision gradients from high-skewness samples are harmful to target classification.
    Supported only by a random-freezing pilot experiment (50% random freeze does not hurt); the connection to skewness is assumed rather than derived (Section I, V.C).
  • domain assumption Additive Noise Model (ANM) from [35] is a valid way to infer causality between skewness and gradient conflict.
    Used in Section V.C; ANM fit scores 0.16 vs 0.25 are reported without confidence intervals or calibration, so the causal direction claim is weak.
  • domain assumption Pseudo-labels in the second stage are reliable enough that saliency maps computed from them remain useful, with a fixed mean 0.1 Gaussian perturbation covering the noise.
    Section V.D; the 0.1 mean is measured on one task and assumed transferable.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Locating and Mitigating Gradient Conflicts in Point Cloud Domain Adaptation via Saliency Map Skewness." pith.science (2026). https://pith.science/paper/6ACVA4HI

@misc{pith2026250415796,
  author       = {Pith},
  title        = {Pith review of: Locating and Mitigating Gradient Conflicts in Point Cloud Domain Adaptation via Saliency Map Skewness},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/6ACVA4HI}},
  note         = {Machine review of arXiv:2504.15796}
}
read the original abstract

Object classification models utilizing point cloud data are fundamental for 3D media understanding, yet they often struggle with unseen or out-of-distribution (OOD) scenarios. Existing point cloud unsupervised domain adaptation (UDA) methods typically employ a multi-task learning (MTL) framework that combines primary classification tasks with auxiliary self-supervision tasks to bridge the gap between cross-domain feature distributions. However, our further experiments demonstrate that not all gradients from self-supervision tasks are beneficial and some may negatively impact the classification performance. In this paper, we propose a novel solution, termed Saliency Map-based Data Sampling Block (SM-DSB), to mitigate these gradient conflicts. Specifically, our method designs a new scoring mechanism based on the skewness of 3D saliency maps to estimate gradient conflicts without requiring target labels. Leveraging this, we develop a sample selection strategy that dynamically filters out samples whose self-supervision gradients are not beneficial for the classification. Our approach is scalable, introducing modest computational overhead, and can be integrated into all the point cloud UDA MTL frameworks. Extensive evaluations demonstrate that our method outperforms state-of-the-art approaches. In addition, we provide a new perspective on understanding the UDA problem through back-propagation analysis.

Figures

Figures reproduced from arXiv: 2504.15796 by the authors.

Figure 1
Figure 1. The left figure illustrates the design principle of our methods: we first use saliency map skewness to estimate the [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Overview of our SM-DSB plugged into the MTL framework. The measurer computes the skewness of instance-level [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Confusion matrices of classifying testing samples on S* [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figures from the paper (8 more)
Figure 4
Figure 4. Figure 4: Conflict inconsistency evaluated on S→M. The figure on the left shows the cosine similarity between aggregated gradients across tasks and the oracle gradients. The figure on the right represents the cosine similarity of the inter-task gradient between the self-supervis…
Figure 5
Figure 5. Figure 5: On the left is a table with a high skewness score, [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]
Figure 6
Figure 6. Figure 6: Loss of self-supervision task and classification task. [PITH_FULL_IMAGE:figures/full_fig_p009_6.png]
Figure 8
Figure 8. Figure 8: The target accuracy under different hyper-parameter [PITH_FULL_IMAGE:figures/full_fig_p009_8.png]
Figure 7
Figure 7. Figure 7: Skewness score discrepancy in the target domain. The [PITH_FULL_IMAGE:figures/full_fig_p009_7.png]
Figure 9
Figure 9. Figure 9: Conflict inconsistency evaluated on S→M. The figure on the left shows the cosine similarity between the gradient sum of the self-supervision task and classification task and the standard gradient. The figure on the right represents the cosine similarity between the gra…
Figure 10
Figure 10. Figure 10: Validation accuracy of different β setting evaluated on M→ S ∗ (left) and S→M(right). F. Quality Analysis We utilize t-SNE to visualize the feature distribution on the target domain of the UDA task S* → M of the baseline without and with our sampling block in [PITH_F…
Figure 11
Figure 11. Figure 11: The t-SNE visualization of feature distribution on the [PITH_FULL_IMAGE:figures/full_fig_p010_11.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

39 extracted references · 26 canonical work pages

  1. [1]

    Pointnet: Deep learning on point sets for 3d classification and segmentation,

    C. R. Qi, H. Su, K. Mo, and L. J. Guibas, “Pointnet: Deep learning on point sets for 3d classification and segmentation,” in Proceedings of the IEEE conference on computer vision and pattern recognition , 2017, pp. 652–660

    2017

  2. [2]

    Pointnet++: Deep hierarchical feature learning on point sets in a metric space,

    C. R. Qi, L. Yi, H. Su, and L. J. Guibas, “Pointnet++: Deep hierarchical feature learning on point sets in a metric space,” Advances in neural information processing systems , vol. 30, 2017

    2017

  3. [3]

    Dynamic graph cnn for learning on point clouds,

    Y . Wang, Y . Sun, Z. Liu, S. E. Sarma, M. M. Bronstein, and J. M. Solomon, “Dynamic graph cnn for learning on point clouds,” Acm Transactions On Graphics (tog) , vol. 38, no. 5, pp. 1–12, 2019

    2019

  4. [4]

    Scannet: Richly-annotated 3d reconstructions of indoor scenes,

    A. Dai, A. X. Chang, M. Savva, M. Halber, T. Funkhouser, and M. Nießner, “Scannet: Richly-annotated 3d reconstructions of indoor scenes,” in Proceedings of the IEEE conference on computer vision and pattern recognition, 2017, pp. 5828–5839

    2017

  5. [5]

    Modelnet: Towards a datacenter emulation environment,

    K. V . Vishwanath, D. Gupta, A. Vahdat, and K. Yocum, “Modelnet: Towards a datacenter emulation environment,” in 2009 IEEE Ninth International Conference on Peer-to-Peer Computing . IEEE, 2009, pp. 81–82

    work page 2009

  6. [6]

    Shapenet: An information- rich 3d model repository,

    A. X. Chang, T. Funkhouser, L. Guibas, P. Hanrahan, Q. Huang, Z. Li, S. Savarese, M. Savva, S. Song, H. Su et al., “Shapenet: An information- rich 3d model repository,” arXiv preprint arXiv:1512.03012 , 2015

    arXiv 2015

  7. [7]

    Domain adaptation on point clouds via geometry-aware implicits,

    Y . Shen, Y . Yang, M. Yan, H. Wang, Y . Zheng, and L. J. Guibas, “Domain adaptation on point clouds via geometry-aware implicits,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2022, pp. 7223–7232

    work page 2022

  8. [8]

    Joint supervised and self-supervised learning for 3d real world challenges,

    A. Alliegro, D. Boscaini, and T. Tommasi, “Joint supervised and self-supervised learning for 3d real world challenges,” in 2020 25th International Conference on Pattern Recognition (ICPR) . IEEE, 2021, pp. 6718–6725

    work page 2020

Show all 39 references
  1. [9]

    Point cloud domain adaptation via masked local 3d structure predic- tion,

    H. Liang, H. Fan, Z. Fan, Y . Wang, T. Chen, Y . Cheng, and Z. Wang, “Point cloud domain adaptation via masked local 3d structure predic- tion,” in Computer Vision–ECCV 2022: 17th European Conference, Tel Aviv, Israel, October 23–27, 2022, Proceedings, Part III . Springer, 202...

    2022

  2. [10]

    Geometry-aware self-training for unsupervised domain adaptation on object point clouds,

    L. Zou, H. Tang, K. Chen, and K. Jia, “Geometry-aware self-training for unsupervised domain adaptation on object point clouds,” in Proceedings of the IEEE/CVF International Conference on Computer Vision , 2021, pp. 6403–6412

    2021

  3. [11]

    Gra- dient surgery for multi-task learning,

    T. Yu, S. Kumar, A. Gupta, S. Levine, K. Hausman, and C. Finn, “Gra- dient surgery for multi-task learning,” Advances in Neural Information Processing Systems, vol. 33, pp. 5824–5836, 2020. 11

    2020

  4. [12]

    Metabalance: improving multi-task recommendations via adapting gradient magni- tudes of auxiliary tasks,

    Y . He, X. Feng, C. Cheng, G. Ji, Y . Guo, and J. Caverlee, “Metabalance: improving multi-task recommendations via adapting gradient magni- tudes of auxiliary tasks,” in Proceedings of the ACM Web Conference 2022, 2022, pp. 2205–2215

    2022

  5. [13]

    Shortcut learning in deep neural networks,

    R. Geirhos, J.-H. Jacobsen, C. Michaelis, R. Zemel, W. Brendel, M. Bethge, and F. A. Wichmann, “Shortcut learning in deep neural networks,” Nature Machine Intelligence , vol. 2, no. 11, pp. 665–673, 2020

    2020

  6. [14]

    Self-distillation for unsupervised 3d domain adaptation,

    A. Cardace, R. Spezialetti, P. Z. Ramirez, S. Salti, and L. Di Stefano, “Self-distillation for unsupervised 3d domain adaptation,” in Proceed- ings of the IEEE/CVF Winter Conference on Applications of Computer Vision, 2023, pp. 4166–4177

    2023

  7. [15]

    Self-supervised learning for domain adaptation on point clouds,

    I. Achituve, H. Maron, and G. Chechik, “Self-supervised learning for domain adaptation on point clouds,” in Proceedings of the IEEE/CVF winter conference on applications of computer vision , 2021, pp. 123– 133

    2021

  8. [16]

    Geometry-aware self-training for unsupervised domain adaptationon object point clouds

    L. Zou, H. Tang, K. Chen, and K. Jia, “Geometry-aware self-training for unsupervised domain adaptationon object point clouds.” [Online]. Available: http://arxiv.org/abs/2108.09169

    arXiv

  9. [17]

    Self-supervised global-local structure modeling for point cloud domain adaptation with reliable voted pseudo labels,

    H. Fan, X. Chang, W. Zhang, Y . Cheng, Y . Sun, and M. Kankanhalli, “Self-supervised global-local structure modeling for point cloud domain adaptation with reliable voted pseudo labels,” in 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) . IEEE, pp. ...

    2022

  10. [18]

    Point cloud domain adaptation via masked local 3d structure prediction,

    H. Liang, H. Fan, Z. Fan, Y . Wang, T. Chen, Y . Cheng, and Z. Wang, “Point cloud domain adaptation via masked local 3d structure prediction,” in Computer Vision – ECCV 2022 , S. Avidan, G. Brostow, M. Ciss ´e, G. M. Farinella, and T. Hassner, Eds. Springer Nature Switzerland,...

    2022 doi

  11. [19]

    A survey of data optimization for problems in computer vision datasets

    Z. Wan, Z. Wang, C. Chung, and Z. Wang, “A survey of data optimization for problems in computer vision datasets.” [Online]. Available: http://arxiv.org/abs/2210.11717

    arXiv

  12. [20]

    A review of instance selection methods,

    J. A. Olvera-L ´opez, J. A. Carrasco-Ochoa, J. F. Mart ´ınez-Trinidad, and J. Kittler, “A review of instance selection methods,” vol. 34, no. 2, pp. 133–143. [Online]. Available: http://link.springer.com/10. 1007/s10462-010-9165-y

  13. [21]

    C4. 5, class imbalance, and cost sensitivity: why under-sampling beats over-sampling,

    C. Drummond, R. C. Holte et al. , “C4. 5, class imbalance, and cost sensitivity: why under-sampling beats over-sampling,” in Workshop on learning from imbalanced datasets II , vol. 11, 2003, pp. 1–8

    2003

  14. [22]

    Crssc: salvage reusable samples from noisy data for robust learning,

    Z. Sun, X.-S. Hua, Y . Yao, X.-S. Wei, G. Hu, and J. Zhang, “Crssc: salvage reusable samples from noisy data for robust learning,” in Proceedings of the 28th ACM International Conference on Multimedia , 2020, pp. 92–101

    2020

  15. [23]

    Distribution alignment: A unified framework for long-tail visual recognition,

    S. Zhang, Z. Li, S. Yan, X. He, and J. Sun, “Distribution alignment: A unified framework for long-tail visual recognition,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , 2021, pp. 2361–2370

    2021

  16. [24]

    Classification with noisy labels by importance reweighting,

    T. Liu and D. Tao, “Classification with noisy labels by importance reweighting,” IEEE Transactions on pattern analysis and machine intelligence, vol. 38, no. 3, pp. 447–461, 2015

    2015

  17. [25]

    Dualgraph: A graph-based method for reasoning about label noise,

    H. Zhang, X. Xing, and L. Liu, “Dualgraph: A graph-based method for reasoning about label noise,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2021, pp. 9654–9663

    2021

  18. [26]

    Conflict-averse gradi- ent descent for multi-task learning,

    B. Liu, X. Liu, X. Jin, P. Stone, and Q. Liu, “Conflict-averse gradi- ent descent for multi-task learning,” Advances in Neural Information Processing Systems, vol. 34, pp. 18 878–18 890, 2021

    2021

  19. [27]

    A model- agnostic approach to mitigate gradient interference for multi-task learn- ing,

    H. Chai, Z. Yin, Y . Ding, L. Liu, B. Fang, and Q. Liao, “A model- agnostic approach to mitigate gradient interference for multi-task learn- ing,” IEEE Transactions on Cybernetics , 2022

    2022

  20. [28]

    Self-supervised deep learning on point clouds by reconstructing space,

    J. Sauder and B. Sievers, “Self-supervised deep learning on point clouds by reconstructing space,” Advances in Neural Information Processing Systems, vol. 32, 2019

    2019

  21. [29]

    Pointcloud saliency maps,

    T. Zheng, C. Chen, J. Yuan, B. Li, and K. Ren, “Pointcloud saliency maps,” in Proceedings of the IEEE/CVF International Conference on Computer Vision, 2019, pp. 1598–1606

    2019

  22. [30]

    3d shapenets: A deep representation for volumetric shapes,

    Z. Wu, S. Song, A. Khosla, F. Yu, L. Zhang, X. Tang, and J. Xiao, “3d shapenets: A deep representation for volumetric shapes,” in Proceedings of the IEEE conference on computer vision and pattern recognition , 2015, pp. 1912–1920

    2015

  23. [31]

    Convolutional neural networks on surfaces via seamless toric covers

    H. Maron, M. Galun, N. Aigerman, M. Trope, N. Dym, E. Yumer, V . G. Kim, and Y . Lipman, “Convolutional neural networks on surfaces via seamless toric covers.” ACM Trans. Graph. , vol. 36, no. 4, pp. 71–1, 2017

    2017

  24. [32]

    Self-distillation for unsupervised 3d domain adaptation,

    A. Cardace, R. Spezialetti, P. Z. Ramirez, S. Salti, and L. D. Stefano, “Self-distillation for unsupervised 3d domain adaptation,” in 2023 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) . IEEE, pp. 4155–4166. [Online]. Available: https: //ieeexplore.ieee....

    2023

  25. [33]

    Pointdan: A multi-scale 3d domain adaption network for point cloud representation,

    C. Qin, H. You, L. Wang, C.-C. J. Kuo, and Y . Fu, “Pointdan: A multi-scale 3d domain adaption network for point cloud representation,” Advances in Neural Information Processing Systems , vol. 32, 2019

    2019

  26. [34]

    Dfan: Dual-branch feature alignment network for domain adaptation on point clouds,

    L. Shi, Z. Yuan, M. Cheng, Y . Chen, and C. Wang, “Dfan: Dual-branch feature alignment network for domain adaptation on point clouds,” IEEE Transactions on Geoscience and Remote Sensing , vol. 60, pp. 1–12, 2022

    2022

  27. [35]

    Nonlinear causal discovery with additive noise models,

    P. Hoyer, D. Janzing, J. M. Mooij, J. Peters, and B. Sch ¨olkopf, “Nonlinear causal discovery with additive noise models,” Advances in neural information processing systems , vol. 21, 2008

    2008

  28. [36]

    Interpreting hidden seman- tics in the intermediate layers of 3d point cloud classification neural network,

    W. Liu, M. Liu, S. Zheng, and C. Wang, “Interpreting hidden seman- tics in the intermediate layers of 3d point cloud classification neural network,” arXiv preprint arXiv:2303.06652 , 2023

    2023 arXiv

  29. [37]

    Multi-task learning using uncer- tainty to weigh losses for scene geometry and semantics,

    A. Kendall, Y . Gal, and R. Cipolla, “Multi-task learning using uncer- tainty to weigh losses for scene geometry and semantics,” inProceedings of the IEEE conference on computer vision and pattern recognition , 2018, pp. 7482–7491

    2018

  30. [38]

    Loss-balanced task weighting to reduce negative transfer in multi-task learning,

    S. Liu, Y . Liang, and A. Gitter, “Loss-balanced task weighting to reduce negative transfer in multi-task learning,” in Proceedings of the AAAI conference on artificial intelligence , vol. 33, no. 01, 2019, pp. 9977– 9978

    2019

  31. [39]

    Dynamic task prioritization for multitask learning,

    M. Guo, A. Haque, D.-A. Huang, S. Yeung, and L. Fei-Fei, “Dynamic task prioritization for multitask learning,” in Proceedings of the Euro- pean conference on computer vision (ECCV) , 2018, pp. 270–287

    2018

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.