Pith. sign in

REVIEW 3 major objections 5 minor 48 references

Data Pruning Can Do More: A Comprehensive Data Pruning Approach for Object Re-identification

T0 review · 3 major / 5 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 Averaging a sample's logits over 12 training epochs and taking the entropy of that averaged soft label yields a single score that prunes redundant re-identification images, corrects mislabeled ones, and flags outliers.

desk verdict Useful, practical ReID data pruning with a real but fixable flaw: the logit-averaging metric overstates its use of training history. read the letter →

arxiv 2412.10091 v1 pith:UZMVPSH4 submitted 2024-12-13 cs.CV cs.AI

Zi Yang , Haojin Yang , Soumajit Majumder , Jorge Cardoso , Guillermo Gallego This is my paper · ORCID
classification cs.CVcs.AI
keywords datapruningobjectre-identificationsoftlabelentropylogittrajectorycorrectionoutlierdetectionsampleimportancetrainingdynamics
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 tries to establish that a single number, the entropy of a sample's time-averaged soft label, captures how important that sample is for training an object re-identification (ReID) model, and that the same soft label can repair mislabeled images and remove outliers. The method records classification logits at every epoch, averages them over the first 12 epochs, and applies softmax to get a soft label; its entropy ranks samples from easy to hard, and low-entropy samples can be pruned without hurting accuracy. On three ReID benchmarks the paper reports pruning 35% of VeRi, 30% of MSMT17, and 5% of Market1501 training data with under 0.1% loss in mean accuracy, while using only 10% of the training budget to compute the scores. The authors also show the same score works on two image-classification datasets, suggesting the mechanism generalizes beyond ReID.

What carries the argument

The time-averaged soft label $\tilde{y} = \sigma\!\bigl(\tfrac{1}{T}\sum_{t=1}^{T} z^{(t)}(x)\bigr)$ — the softmax of the mean logit vector across $T=12$ training epochs — is the central object of the paper. Its entropy $H(\tilde{y})$ provides the importance ranking, its argmax provides the corrected label for mislabeled samples, and its maximum class score acts as the outlier detector when it falls below a threshold $\delta$. The 12-epoch window is what makes the approach cheap: importance is estimated in 10% of the usual 120-epoch training budget, a tenfold reduction in scoring cost relative to methods that need a full training run.

What would settle it

Compute the mean logit magnitude per epoch over the 12-epoch estimation window on MSMT17; if the scale drifts substantially (e.g., more than a factor of 2 between early and late epochs), recompute the soft labels after standardizing each epoch's logits to unit variance before averaging, and compare the resulting pruning ranking and accuracy-versus-pruning curves with the raw-average version. A material difference would falsify the claim that the raw average reflects the full training history.

Watch

Extended reading notes

Core claim

The central claim is that the entropy of the time-averaged soft label is a more accurate and cheaper measure of sample importance than single-epoch error norms, forgetting events, or feature-space distances. For each image the paper records the classification-head logits at every epoch, averages them across $T=12$ epochs, and applies softmax to obtain $\tilde{y} = \sigma\bigl(\tfrac{1}{T}\sum_{t=1}^{T} z^{(t)}(x)\bigr)$; the importance score is the entropy $H(\tilde{y}) = -\sum_i p_i \log_2 p_i$. High-entropy samples are treated as hard and valuable; low-entropy samples can be pruned. The same soft label drives purification: if $\arg\max(\tilde{y}) \neq y$ the image is re-labeled as $\arg\max(\tilde{y})$, and if $\max(\tilde{y}) \leq \delta = 10\%$ it is removed as an outlier. Integrated, the procedure removes easy samples, repairs mislabeled ones, and deletes outliers, yielding pruned datasets that train faster with under 0.1% mean-accuracy loss on three ReID benchmarks and, as an extension, on two classification benchmarks.

Load-bearing premise

The load-bearing premise is that raw logits from different epochs can be averaged directly without rescaling: if logit magnitudes drift during training, the average is dominated by the last epochs and the entropy ranking no longer represents the whole training history.

Editorial extensions

If this is right

  • On VeRi, MSMT17, and Market1501, the full pipeline (pruning plus label correction and outlier removal) removes 35%, 30%, and 5% of training samples respectively — with proportionally less training time — while the mean of rank-1 accuracy and mAP stays within 0.1% of the full-data model.
  • A ResNet50-derived importance ranking transfers to ResNet101 and ViT-B/16, so the pruning step does not need to be repeated when the architecture changes.
  • Injected label noise of 10% to 50% is largely neutralized by combining label correction (re-label to the argmax of the averaged soft label) with outlier removal (max soft-label score below the threshold), outperforming the AUM baseline on the same noisy settings.
  • Because the score is computed in 12 epochs rather than a full 120-epoch run, importance estimation costs about 15.8 minutes instead of 2.6 to 5.2 hours on MSMT17, a roughly tenfold saving.
  • The same importance score works for CIFAR-100 and CUB-200-2011 without pretrained backbones, indicating the mechanism is not specific to re-identification.

Reading between the lines

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

  • If logit magnitudes drift across epochs, the raw average in Eq. 1 is dominated by later epochs; a variant that standardizes each epoch's logits before averaging would test whether the claimed use of the full training dynamics is real, and would change the ranking only if the drift is large.
  • The entropy of the time-averaged soft label is effectively an uncertainty measure over the training trajectory, so the same quantity could serve as a curriculum ordering or sample-weighting signal during training, not just a one-shot pruning score.
  • Because the soft label is a full distribution over identities, the method could be repurposed to detect identity-level annotation errors or quantify inter-identity similarity, which the paper does not explore.
  • The outlier threshold is fixed at 10% of the maximum class score; a data-driven choice based on the distribution of max scores would make the method applicable when the noise rate is unknown.
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

3 major / 5 minor

Summary. The paper proposes a data pruning and purification framework for object re-identification. For each training sample, it records the logits of the classification head over T=12 training epochs, averages them, applies softmax to form a 'soft label', and uses the entropy of this soft label as the importance score. It then prunes low-entropy (easy) samples, corrects mislabeled samples when argmax of the soft label disagrees with the assigned label, and removes outliers whose maximum soft-label score is below a threshold delta. Experiments on Market1501, MSMT17, and VeRi with ResNet50 (and generalization to ResNet101/ViT) report that 5%, 30%, and 35% of samples can be removed with negligible (<0.1%) loss in accuracy, and the method also improves accuracy on synthetically noised datasets. The code and sample lists are publicly available.

Significance. If the empirical claims hold, the paper provides a practical, architecture-agnostic data pruning tool for ReID, a task where data pruning has been little studied. The method's estimated low cost (12 epochs, no multi-model ensembling) and the release of sample importance, mislabel, and outlier lists are concrete contributions. The paper includes a useful assessment of existing pruning baselines on ReID and demonstrates cross-architecture transfer of the learned ranking. However, the conceptual novelty rests on the untested claim that the soft label genuinely aggregates the full training trajectory, and the key comparisons contain a loss-function confound; these issues prevent the current version from fully supporting the stated claims.

major comments (3)
  1. [Sec. 5.3 / Appendix A.2] The comparison between the proposed score and EL2N is confounded by the training loss. The proposed score is computed from logits of a model trained with cross-entropy plus triplet loss, while the EL2N baselines use cross-entropy only, as stated in Appendix A.2. With metric loss, the model produces different logit geometry, and the better pruning curves in Fig. 4 could reflect a better-trained model rather than a better importance metric. Please rerun EL2N with the same CE+triplet loss (or rerun the proposed method with CE only) and confirm that the gap remains.
  2. [Sec. 3.3, Eq. 1, Fig. 2, Figs. 9-11] Eq. 1 averages raw logits across epochs without any per-epoch normalization. The paper's own Fig. 2 shows logit magnitudes growing from roughly 0 to about 7 over 13 epochs. Since softmax is monotone, the averaged logits are dominated by the later high-magnitude epochs, so the 'soft label' is close to the final-epoch prediction. Consequently, the claimed benefit of 'fully exploiting the logit trajectory' is not established. The ablations in Figs. 10-11 compare accumulation frequencies (every 1/2/4/6 epochs) but are also explained by scale: adding more epochs mostly adds later large logits. The 12-epoch result in Fig. 9 is likewise not evidence for trajectory use. Please normalize logits per epoch (e.g., standardize or use temperature) before averaging, or directly compare against a 'last-epoch only' baseline to show what the accumulation adds.
  3. [Fig. 4 / Sec. 5.3.1] The headline claim of removing 35%, 30%, and 5% of samples with <0.1% accuracy loss is presented without error bars in Fig. 4, although the caption reports means over four seeds. Without measures of variance on the pruning curves (and a paired comparison to the no-pruning baseline), the reader cannot assess whether the differences are within noise. Please add error bars or confidence intervals to all main pruning figures and report the per-seed accuracy differences against the no-pruning baseline.
minor comments (5)
  1. [Appendix D] There is a typo at the end of Appendix D.1: 'e.g. 10%, 20%.,' should read 'e.g., 10%, 20%.'
  2. [Sec. 5.3.1] In the 'Differences in Datasets' paragraph, 'MSMT' is used once instead of 'MSMT17'; please make the dataset name consistent.
  3. [Appendix A.2] The comparison with supervised prototypes replaces the original self-supervised learning with a supervised variant; this deviation should be acknowledged in the main text of Sec. 5.3.1 as well, not only in the appendix.
  4. [Sec. 3.3] The phrase 'a logit value, presented on a logarithmic scale, represents the network's predicted probability of a particular class' is imprecise; logits are not generally log-probabilities. Please rephrase to avoid confusion.
  5. [Abstract / Table 1] The abstract says the approach reduces importance-score estimation cost by 10 times, but Table 1 shows roughly 10x versus forgetting/prototypes and roughly 20x versus EL2N(20 models). Please clarify the reference point for the 10x claim.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the importance metric and purification components are validated on hold-out retraining accuracy and synthetic label noise with known ground truth.

full rationale

The paper's derivation chain is empirical rather than circular. Equation (1) defines the soft label as the softmax of the average raw logits across training epochs, and Equation (2) defines the importance score as the entropy of that soft label; neither equation encodes the target pruning result. The pruning ranking is validated by retraining models from scratch on pruned datasets and comparing against random pruning and established baselines such as EL2N, forgetting scores, and supervised prototypes, so the metric is not a fitted parameter renamed as a prediction. The label-correction and outlier-removal components are validated on synthetically corrupted labels with known ground truth, which provides an external check on the assumption that argmax of the averaged soft label recovers the true label. The hyperparameters T and delta are selected after ablations on the same datasets, but the sensitivity analyses show stability across a range of values, and this is standard empirical tuning rather than a circular reduction. There are no load-bearing self-citations, no imported uniqueness theorems, and no known results renamed as new contributions. The only substantive concern, namely that raw logits may grow in magnitude during training so that later epochs dominate the unweighted average in Equation (1), is a correctness or robustness risk, not a circularity, because the subsequent evaluation is independent of how the soft label is justified.

Assumptions & free parameters 2 free parameters · 4 assumptions · 0 invented entities

The central claim depends on two hand-set hyperparameters (T=12, delta=10%), both shown to be robust but not derived, and on two domain assumptions: that difficulty (entropy) is the right pruning criterion, and that the averaged-soft-label argmax identifies true labels. The logit-averaging assumption (Eq. 1) is the least examined, since no normalization for scale drift is attempted. No invented entities are introduced.

free parameters (2)
  • Outlier removal threshold delta = 10% (max soft-label score)
    Samples with max(y~) <= delta are removed as outliers (Sec 4). Set to 10% on all ReID datasets; sensitivity analysis (Fig 12) shows robust performance for delta in 5-15%, so it is a hand-set threshold that gates how much data is discarded, not derived from first principles.
  • Soft-label computation epoch T = 12 epochs
    Number of epochs used to accumulate logits in Eq. 1. Selected because performance stabilizes after 12 epochs (Fig 9), which is 10% of the 120-epoch training budget. T affects both importance ranking and label correction; it is tuned on MSMT17 and applied across datasets.
assumptions (4)
  • domain assumption Higher sample difficulty implies higher importance for training, so entropy can serve as an importance surrogate
    The method prunes low-entropy (easy) samples based on the prior pruning literature (Sec 3.2, citing Toneva et al. 2018 and Paul et al. 2021). This underlies the entire pruning step.
  • domain assumption Raw logits across epochs are directly averageable without normalization
    Eq. 1 averages pre-softmax logits from all epochs. The paper does not discuss logit scale drift across training; if scales grow, the average is not a faithful representation of the training trajectory.
  • ad hoc to paper The argmax of the averaged soft label is the true label of a mislabeled sample
    Sec 4 corrects labels by setting y = argmax(y~). This is validated only on synthetically corrupted labels (Sec 5.4); on real ReID label noise, the correction accuracy is asserted, not measured against ground truth.
  • standard math Standard softmax and entropy definitions
    Eqs. 1 and 2 use standard softmax and Shannon entropy with log base 2; no special mathematical assumptions are introduced.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Data Pruning Can Do More: A Comprehensive Data Pruning Approach for Object Re-identification." pith.science (2026). https://pith.science/paper/UZMVPSH4

@misc{pith2026241210091,
  author       = {Pith},
  title        = {Pith review of: Data Pruning Can Do More: A Comprehensive Data Pruning Approach for Object Re-identification},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/UZMVPSH4}},
  note         = {Machine review of arXiv:2412.10091}
}
read the original abstract

Previous studies have demonstrated that not each sample in a dataset is of equal importance during training. Data pruning aims to remove less important or informative samples while still achieving comparable results as training on the original (untruncated) dataset, thereby reducing storage and training costs. However, the majority of data pruning methods are applied to image classification tasks. To our knowledge, this work is the first to explore the feasibility of these pruning methods applied to object re-identification (ReID) tasks, while also presenting a more comprehensive data pruning approach. By fully leveraging the logit history during training, our approach offers a more accurate and comprehensive metric for quantifying sample importance, as well as correcting mislabeled samples and recognizing outliers. Furthermore, our approach is highly efficient, reducing the cost of importance score estimation by 10 times compared to existing methods. Our approach is a plug-and-play, architecture-agnostic framework that can eliminate/reduce 35%, 30%, and 5% of samples/training time on the VeRi, MSMT17 and Market1501 datasets, respectively, with negligible loss in accuracy (< 0.1%). The lists of important, mislabeled, and outlier samples from these ReID datasets are available at https://github.com/Zi-Y/data-pruning-reid.

Figures

Figures reproduced from arXiv: 2412.10091 by the authors.

Figure 1
Figure 1. (a) The workflow of data pruning. (b) Our data pruning approach not only identifies [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Logit trajectories for three samples (i.e., the evolution of the log probabilities of each sample [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Illustration of the generated soft labels averaged over 12 training epochs for different sample types [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (15 more)
Figure 4
Figure 4. Figure 4: Data pruning on ReID datasets. We report the mean of Rank1 and mAP on 3 ReID datasets [PITH_FULL_IMAGE:figures/full_fig_p009_4.png]
Figure 5
Figure 5. Figure 5: Generalization performance. We train a ResNet101 and a ViT model using the [PITH_FULL_IMAGE:figures/full_fig_p010_5.png]
Figure 6
Figure 6. Figure 6: Data pruning on classification datasets. We report the mean of final test accuracy on two clas [PITH_FULL_IMAGE:figures/full_fig_p010_6.png]
Figure 7
Figure 7. Figure 7: Data purification on noisy datasets. We report accuracy under different random noise ratios. Each curve represents the mean accuracy over four independent runs. From [PITH_FULL_IMAGE:figures/full_fig_p011_7.png]
Figure 8
Figure 8. Figure 8: Data pruning with data purification. Accuracy is achieved by training on the pruned dataset [PITH_FULL_IMAGE:figures/full_fig_p011_8.png]
Figure 9
Figure 9. Figure 9: Impact of score computation epoch T. Model performance achieved by training on 85% training data com￾prised of examples with maximum for￾getting, EL2N, and our scores com￾puted at different epochs. Impact of Score Computation Epoch T. We investigate how early in traini…
Figure 10
Figure 10. Figure 10: Ablation study of logit ac￾cumulation for data pruning. Model performance when trained with the importance scores computed using dif￾ferent frequencies of logit accumula￾tion. 10 20 30 40 50 Random Noise Ratio (%) 0 20 40 60 80 Avg. accuracy (%): (Rank1 + mAP)/2 Marke…
Figure 12
Figure 12. Figure 12: Impact of outlier removal threshold δ. Model performance under noise ratio 10%, 20% and 30%. 6 Discussion and Conclusion In this work, we have addressed two issues of ReID datasets (less informative samples and noise) by proposing a plug-and-play architecture-agnostic…
Figure 13
Figure 13. Figure 13: A general supervised ReID workflow. A.2 Existing Methods E2LN. We use the same model configuration and settings as in our approach, except for the loss function. When calculating the EL2N score, we strictly follow its procedure and definition (Paul et al., 2021), i.e.…
Figure 14
Figure 14. Figure 14: Examples of different outliers from MSMT17. [PITH_FULL_IMAGE:figures/full_fig_p019_14.png]
Figure 15
Figure 15. Figure 15: Accuracy achieved by training on the Market1501 dataset under different noise ratios [PITH_FULL_IMAGE:figures/full_fig_p019_15.png]
Figure 16
Figure 16. Figure 16: Samples from Market1501 sorted from easy (the first column) to hard (the last column) based on our proposed importance scores. Images in each row belong to the same identity. 20 [PITH_FULL_IMAGE:figures/full_fig_p020_16.png]
Figure 17
Figure 17. Figure 17: Samples from MSMT17 sorted from easy (the first column) to hard (the last column) based on our proposed importance scores. Images in each row belong to the same identity. 21 [PITH_FULL_IMAGE:figures/full_fig_p021_17.png]
Figure 18
Figure 18. Figure 18: Samples from VeRi sorted from easy (the first column) to hard (the last column) based on our proposed importance scores. Images in each row belong to the same identity. 22 [PITH_FULL_IMAGE:figures/full_fig_p022_18.png]
Figure 19
Figure 19. Figure 19: Samples from CIFAR-100 sorted from easy (the first column) to hard (the last column) based on our proposed importance scores. Images in each row belong to the same identity. 23 [PITH_FULL_IMAGE:figures/full_fig_p023_19.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

48 extracted references · 29 canonical work pages

  1. [1]

    A representative consumer theory of the logit model

    Simon P Anderson, Andr \'e De Palma, and J-F Thisse. A representative consumer theory of the logit model. International Economic Review, pp.\ 461--466, 1988

    work page 1988

  2. [2]

    A survey of approaches and trends in person re-identification

    Apurva Bedagkar-Gala and Shishir K Shah. A survey of approaches and trends in person re-identification. Image and vision computing, 32 0 (4): 0 270--286, 2014

    work page 2014

  3. [3]

    A unifying mutual information view of metric learning: cross-entropy vs

    Malik Boudiaf, J \'e r \^o me Rony, Imtiaz Masud Ziko, Eric Granger, Marco Pedersoli, Pablo Piantanida, and Ismail Ben Ayed . A unifying mutual information view of metric learning: cross-entropy vs. pairwise losses. In European Conference on Computer Vision, pp.\ 548--564. Springer, 2020

    work page 2020

  4. [4]

    Beyond triplet loss: a deep quadruplet network for person re-identification

    Weihua Chen, Xiaotang Chen, Jianguo Zhang, and Kaiqi Huang. Beyond triplet loss: a deep quadruplet network for person re-identification. In Proceedings of the IEEE conference on computer vision and pattern recognition, pp.\ 403--412, 2017

    work page 2017

  5. [5]

    Imagenet: A large-scale hierarchical image database

    Jia Deng, Wei Dong, Richard Socher, Li-Jia Li, Kai Li, and Li Fei-Fei. Imagenet: A large-scale hierarchical image database. In 2009 IEEE conference on computer vision and pattern recognition, pp.\ 248--255. Ieee, 2009

    2009

  6. [6]

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

    Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Sylvain Gelly, Jakob Uszkoreit, and Neil Houlsby. An image is worth 16x16 words: Transformers for image recognition at scale. ICLR, 2021

    2021

  7. [7]

    Adversarial active learning for deep networks: a margin based approach

    Melanie Ducoffe and Frederic Precioso. Adversarial active learning for deep networks: a margin based approach. arXiv preprint arXiv:1802.09841, 2018

    arXiv 2018

  8. [8]

    What neural networks memorize and why: Discovering the long tail via influence estimation

    Vitaly Feldman and Chiyuan Zhang. What neural networks memorize and why: Discovering the long tail via influence estimation. Advances in Neural Information Processing Systems, 33: 0 2881--2891, 2020

    2020

Show all 48 references
  1. [9]

    Large-scale pre-training for person re-identification with noisy labels

    Dengpan Fu, Dongdong Chen, Hao Yang, Jianmin Bao, Lu Yuan, Lei Zhang, Houqiang Li, Fang Wen, and Dong Chen. Large-scale pre-training for person re-identification with noisy labels. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp.\ 2476-...

    2022

  2. [10]

    Deep residual learning for image recognition

    Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition, pp.\ 770--778, 2016

    2016

  3. [11]

    Fastreid: A pytorch toolbox for general instance re-identification

    Lingxiao He, Xingyu Liao, Wu Liu, Xinchen Liu, Peng Cheng, and Tao Mei. Fastreid: A pytorch toolbox for general instance re-identification. arXiv preprint arXiv:2006.02631, 2020

    2006 arXiv

  4. [12]

    In defense of the triplet loss for person re-identification

    Alexander Hermans, Lucas Beyer, and Bastian Leibe. In defense of the triplet loss for person re-identification. arXiv preprint arXiv:1703.07737, 2017

    2017 arXiv

  5. [13]

    Deep-reid: Deep features and autoencoder assisted image patching strategy for person re-identification in smart cities surveillance

    Samee Ullah Khan, Tanveer Hussain, Amin Ullah, and Sung Wook Baik. Deep-reid: Deep features and autoencoder assisted image patching strategy for person re-identification in smart cities surveillance. Multimedia Tools and Applications, pp.\ 1--22, 2021

    2021

  6. [14]

    Adam: A method for stochastic optimization

    Diederik P Kingma and Jimmy Ba. Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980, 2014

    2014 arXiv

  7. [15]

    Learning multiple layers of features from tiny images, 2009

    Alex Krizhevsky, Geoffrey Hinton, et al. Learning multiple layers of features from tiny images, 2009. URL https://www.cs.toronto.edu/ kriz/learning-features-2009-TR.pdf

    2009

  8. [16]

    Graph sampling based deep metric learning for generalizable person re-identification

    Shengcai Liao and Ling Shao. Graph sampling based deep metric learning for generalizable person re-identification. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp.\ 7359--7368, 2022

    2022

  9. [17]

    A deep learning-based approach to progressive vehicle re-identification for urban surveillance

    Xinchen Liu, Wu Liu, Tao Mei, and Huadong Ma. A deep learning-based approach to progressive vehicle re-identification for urban surveillance. In Computer Vision--ECCV 2016: 14th European Conference, Amsterdam, The Netherlands, October 11-14, 2016, Proceedings, Part II 14, pp.\...

    2016

  10. [18]

    Bag of tricks and a strong baseline for deep person re-identification

    Hao Luo, Youzhi Gu, Xingyu Liao, Shenqi Lai, and Wei Jiang. Bag of tricks and a strong baseline for deep person re-identification. In The IEEE Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, June 2019

    2019

  11. [19]

    Time lags in biological models, volume 27

    Norman MacDonald. Time lags in biological models, volume 27. Springer Science & Business Media, 2013

    2013

  12. [20]

    Pytorch: An imperative style, high-performance deep learning library

    Adam Paszke, Sam Gross, Francisco Massa, Adam Lerer, James Bradbury, Gregory Chanan, Trevor Killeen, Zeming Lin, Natalia Gimelshein, Luca Antiga, Alban Desmaison, Andreas Kopf, Edward Yang, Zachary DeVito, Martin Raison, Alykhan Tejani, Sasank Chilamkurthy, Benoit Steiner, Lu ...

    2019

  13. [21]

    Deep learning on a data diet: Finding important examples early in training

    Mansheej Paul, Surya Ganguli, and Gintare Karolina Dziugaite. Deep learning on a data diet: Finding important examples early in training. Advances in Neural Information Processing Systems, 34: 0 20596--20607, 2021

    2021

  14. [22]

    Identifying mislabeled data using the area under the margin ranking

    Geoff Pleiss, Tianyi Zhang, Ethan Elenberg, and Kilian Q Weinberger. Identifying mislabeled data using the area under the margin ranking. Advances in Neural Information Processing Systems, 33: 0 17044--17056, 2020

    2020

  15. [23]

    Improved person re-identification based on saliency and semantic parsing with deep neural network models

    Rodolfo Quispe and Helio Pedrini. Improved person re-identification based on saliency and semantic parsing with deep neural network models. Image and Vision Computing, 92: 0 103809, 2019

    2019

  16. [24]

    Deep convolutional neural networks for image classification: A comprehensive review

    Waseem Rawat and Zenghui Wang. Deep convolutional neural networks for image classification: A comprehensive review. Neural computation, 29 0 (9): 0 2352--2449, 2017

    2017

  17. [25]

    Facenet: A unified embedding for face recognition and clustering

    Florian Schroff, Dmitry Kalenichenko, and James Philbin. Facenet: A unified embedding for face recognition and clustering. In Proceedings of the IEEE conference on computer vision and pattern recognition, pp.\ 815--823, 2015

    2015

  18. [26]

    Active learning for convolutional neural networks: A core-set approach

    Ozan Sener and Silvio Savarese. Active learning for convolutional neural networks: A core-set approach. arXiv preprint arXiv:1708.00489, 2017

    2017 arXiv

  19. [27]

    Training region-based object detectors with online hard example mining

    Abhinav Shrivastava, Abhinav Gupta, and Ross Girshick. Training region-based object detectors with online hard example mining. In Proceedings of the IEEE conference on computer vision and pattern recognition, pp.\ 761--769, 2016

    2016

  20. [28]

    Compact triplet loss for person re-identification in camera sensor networks

    Tongzhen Si, Zhong Zhang, and Shuang Liu. Compact triplet loss for person re-identification in camera sensor networks. Ad Hoc Networks, 95: 0 101984, 2019

    2019

  21. [29]

    Generalizable person re-identification by domain-invariant mapping network

    Jifei Song, Yongxin Yang, Yi-Zhe Song, Tao Xiang, and Timothy M Hospedales. Generalizable person re-identification by domain-invariant mapping network. In Proceedings of the IEEE/CVF conference on Computer Vision and Pattern Recognition, pp.\ 719--728, 2019

    2019

  22. [30]

    Beyond neural scaling laws: beating power law scaling via data pruning

    Ben Sorscher, Robert Geirhos, Shashank Shekhar, Surya Ganguli, and Ari Morcos. Beyond neural scaling laws: beating power law scaling via data pruning. Advances in Neural Information Processing Systems, 35: 0 19523--19536, 2022

    2022

  23. [31]

    An empirical study of example forgetting during deep neural network learning

    Mariya Toneva, Alessandro Sordoni, Remi Tachet des Combes, Adam Trischler, Yoshua Bengio, and Geoffrey J Gordon. An empirical study of example forgetting during deep neural network learning. arXiv preprint arXiv:1812.05159, 2018

    2018 arXiv

  24. [32]

    The C altech- UCSD birds-200-2011 dataset

    Catherine Wah, Steve Branson, Peter Welinder, Pietro Perona, and Serge Belongie. The C altech- UCSD birds-200-2011 dataset. 2011

    2011

  25. [33]

    A real-time and unsupervised face re-identification system for human-robot interaction

    Yujiang Wang, Jie Shen, Stavros Petridis, and Maja Pantic. A real-time and unsupervised face re-identification system for human-robot interaction. Pattern Recognition Letters, 128: 0 559--568, 2019

    2019

  26. [34]

    Person transfer gan to bridge domain gap for person re-identification

    Longhui Wei, Shiliang Zhang, Wen Gao, and Qi Tian. Person transfer gan to bridge domain gap for person re-identification. In Proceedings of the IEEE conference on computer vision and pattern recognition, pp.\ 79--88, 2018

    2018

  27. [35]

    May i be your personal coach? bringing together person tracking and visual re-identification on a mobile robot

    Tim Wengefeld, Markus Eisenbach, Thanh Q Trinh, and Horst-Michael Gross. May i be your personal coach? bringing together person tracking and visual re-identification on a mobile robot. In Proceedings of ISR 2016: 47st International Symposium on Robotics, pp.\ 1--8. VDE, 2016

    2016

  28. [36]

    A novel deep model with multi-loss and efficient training for person re-identification

    Di Wu, Si-Jia Zheng, Wen-Zheng Bao, Xiao-Ping Zhang, Chang-An Yuan, and De-Shuang Huang. A novel deep model with multi-loss and efficient training for person re-identification. Neurocomputing, 324: 0 69--75, 2019

    2019

  29. [37]

    Person re-identification using deep learning networks: A systematic review

    Ankit Yadav and Dinesh Kumar Vishwakarma. Person re-identification using deep learning networks: A systematic review. arXiv preprint arXiv:2012.13318, 2020

    2012 arXiv

  30. [38]

    Dataset pruning: Reducing training data by examining generalization influence

    Shuo Yang, Zeke Xie, Hanyu Peng, Min Xu, Mingming Sun, and Ping Li. Dataset pruning: Reducing training data by examining generalization influence. arXiv preprint arXiv:2205.09329, 2022

    2022 arXiv

  31. [39]

    Deep learning for person re-identification: A survey and outlook

    Mang Ye, Jianbing Shen, Gaojie Lin, Tao Xiang, Ling Shao, and Steven CH Hoi. Deep learning for person re-identification: A survey and outlook. IEEE transactions on pattern analysis and machine intelligence, 44 0 (6): 0 2872--2893, 2021

    2021

  32. [40]

    Robust person re-identification by modelling feature uncertainty

    Tianyuan Yu, Da Li, Yongxin Yang, Timothy M Hospedales, and Tao Xiang. Robust person re-identification by modelling feature uncertainty. In Proceedings of the IEEE/CVF international conference on computer vision, pp.\ 552--561, 2019

    2019

  33. [41]

    In defense of the triplet loss again: Learning robust person re-identification with fast approximated triplet loss and label distillation

    Ye Yuan, Wuyang Chen, Yang Yang, and Zhangyang Wang. In defense of the triplet loss again: Learning robust person re-identification with fast approximated triplet loss and label distillation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition ...

    2020

  34. [42]

    Person re-identification: A retrospective on domain specific open challenges and future trends

    Asmat Zahra, Nazia Perwaiz, Muhammad Shahzad, and Muhammad Moazam Fraz. Person re-identification: A retrospective on domain specific open challenges and future trends. arXiv preprint arXiv:2202.13121, 2022

    2022 arXiv

  35. [43]

    Relation classification via convolutional deep neural network

    Daojian Zeng, Kang Liu, Siwei Lai, Guangyou Zhou, and Jun Zhao. Relation classification via convolutional deep neural network. In Proceedings of COLING 2014, the 25th international conference on computational linguistics: technical papers, pp.\ 2335--2344, 2014

    2014

  36. [44]

    Decoupled knowledge distillation

    Borui Zhao, Quan Cui, Renjie Song, Yiyu Qiu, and Jiajun Liang. Decoupled knowledge distillation. In Proceedings of the IEEE/CVF Conference on computer vision and pattern recognition, pp.\ 11953--11962, 2022

    2022

  37. [45]

    Scalable person re-identification: A benchmark

    Liang Zheng, Liyue Shen, Lu Tian, Shengjin Wang, Jingdong Wang, and Qi Tian. Scalable person re-identification: A benchmark. In Computer Vision, IEEE International Conference on, 2015

    2015

  38. [46]

    Person re-identification: Past, present and future

    Liang Zheng, Yi Yang, and Alexander G Hauptmann. Person re-identification: Past, present and future. arXiv preprint arXiv:1610.02984, 2016

    2016 arXiv

  39. [47]

    Robust curriculum learning: from clean label detection to noisy label self-correction

    Tianyi Zhou, Shengjie Wang, and Jeff Bilmes. Robust curriculum learning: from clean label detection to noisy label self-correction. In International Conference on Learning Representations, 2020

    2020

  40. [48]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 gl...

Pith tools

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