Pith. sign in

REVIEW 4 major objections 6 minor 39 references

Ambiguity-Restrained Text-Video Representation Learning for Partially Relevant Video Retrieval

T0 review · 4 major / 6 minor · reviewed 2026-08-07 · deepseek-v4-flash

Pith's one-line read Partially relevant video retrieval fails when every unpaired video is treated as a hard negative; this paper detects ambiguous pairs and relaxes their training constraint, reporting state-of-the-art recall on TVR and ActivityNet.

desk verdict A plausible, well-engineered method for treating ambiguous text-video pairs in PRVR, but the central ambiguity-detection mechanism lacks direct validation and the SOTA claim needs variance and external checks. read the letter →

arxiv 2506.07471 v1 pith:WULKKNXC submitted 2025-06-09 cs.CV cs.AI

classification cs.CVcs.AI
keywords partiallyrelevantvideoretrievaltext-videolabelambiguityuncertaintyestimationcontrastivelearningmulti-positivedualtripletmarginlosscross-model
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

Partially relevant video retrieval asks a model to find untrimmed videos in which some segment matches a text query, yet standard training labels every unpaired video as a hard negative. This paper argues that such pair-wise labels are often wrong in spirit: an unpaired video can share the query's context even when it is not the annotated match, so forcing it away from the query corrupts the learned representation. To counter this, ARL first flags ambiguous text-video pairs using two complementary signals—dataset-level uncertainty (how much a query or video overlaps with many other instances) and pair-level similarity (the best matching frame)—then trains with those pairs treated neither as full positives nor as hard negatives, via multi-positive contrastive learning and a dual triplet margin loss. The method extends the same relaxation down to individual frames inside a video and uses two symmetric encoders that exchange their detected ambiguous sets to keep errors from compounding. On TVR and ActivityNet Captions, the reported result is a state-of-the-art retrieval score, with SumR (sum of recall at ranks 1, 5, 10, and 100) gains of 9.3 and 2.3 points over the prior best model in the ResNet/I3D/RoBERTa setup.

What carries the argument

The mechanism that carries the argument is a pair of Label Ambiguity Detection (LAD) modules operating at two granularities. Text-video LAD computes a dataset-wise uncertainty for every query and video (the average similarity of that instance to all instances of the other modality) and a pair-level similarity for each query-video pair (the maximum cosine similarity between the query and the video's frames), then flags a pair as ambiguous when both values exceed thresholds set to the mean of the training distributions. Text-frame LAD applies the same test between a query and the frames of a single untrimmed video. Detected ambiguous sets are fed into a multi-positive supervised contrastive loss and a dual triplet margin loss with a smaller margin for ambiguous samples than for true negatives, and a two-encoder cross-model scheme lets each encoder train on the other's detected ambiguous sets.

What would settle it

Annotate a held-out set of unpaired text-video pairs from TVR or ActivityNet as truly relevant or truly irrelevant; if the pairs ARL flags as ambiguous are mostly judged irrelevant by human raters, or if removing the ambiguity relaxation (training them as hard negatives) does not lower recall, the central claim is falsified.

Watch

Extended reading notes

Core claim

The paper's central claim is that the one-to-one labeling convention used to train PRVR models—one positive video per query, every other video negative—actively harms retrieval because untrimmed videos routinely share context with unpaired queries. It asserts that this ambiguity can be detected online with two complementary metrics, uncertainty and similarity, and that restraining the loss on detected ambiguous pairs instead of treating them as negatives yields better text-video alignment. On the paper's own terms, ARL is the first method to make this ambiguity explicit in PRVR, and it reports state-of-the-art recall on both TVR and ActivityNet Captions, including a 9.3-point SumR improvement over the strongest prior model in the standard feature setting and 9.1- and 3.3-point gains with CLIP-L/14 features.

Load-bearing premise

The method depends on the assumption that a pair it flags as ambiguous—high average overlap plus high pairwise similarity—really is a relevant-but-unlabeled pair, and that the chosen mean thresholds separate those from ordinary negatives.

Editorial extensions

If this is right

  • On TVR and ActivityNet Captions, ARL achieves higher R@1, R@5, R@10, R@100, and SumR than prior PRVR methods under the same features, with the largest reported margin being a 9.3-point SumR gain in the ResNet/I3D/RoBERTa setting.
  • The text-frame ambiguity module lets the model use information from every frame of an untrimmed video, not just the single best-matching frame, without any additional annotation.
  • Cross-model ambiguity detection prevents one model from repeatedly reinforcing its own mistaken ambiguity predictions, so the training signal stays cleaner as training progresses.
  • Because the thresholds are set each epoch from the mean of the training similarity and uncertainty distributions, the method adapts to the model's changing representation rather than relying on fixed cutoffs.

Reading between the lines

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

  • This suggests the same two-signal ambiguity test could be used to audit other pairwise-labeled retrieval datasets: instances with high dataset-level uncertainty and high pair similarity are candidate false negatives, and re-weighting them as soft positives might improve retrieval there too.
  • A natural stress test is to replace the full-dataset uncertainty computation with an online or approximate estimate; if results hold, the method scales to much larger corpora where an exact similarity map over all pairs each epoch is costly.
  • The text-frame ambiguity formulation points toward weakly supervised video moment grounding: the frames that survive the ambiguity test inside a positive video are, in effect, pseudo-labels for the relevant segment.
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 / 6 minor

Summary. The paper proposes Ambiguity-Restrained Representation Learning (ARL) for Partially Relevant Video Retrieval (PRVR). ARL detects ambiguous text-video pairs using two criteria: uncertainty, defined as the average similarity of an instance to all instances in the dataset, and pairwise similarity, defined as the maximum text-frame similarity. Detected ambiguous pairs are then treated neither as hard negatives nor as positives; instead they are incorporated into a multi-positive contrastive loss and a dual triplet margin loss at both text-video and text-frame levels. A cross-model mechanism exchanges detected ambiguous sets between two identical encoders to reduce error propagation. Experiments on TVR and ActivityNet Captions report improved SumR over prior PRVR methods, with ablations showing contributions from each component.

Significance. If the ambiguity-detection mechanism is valid, the paper offers a new and potentially useful perspective on PRVR: pair-wise labeling in retrieval datasets creates ambiguous relations that should be modeled softly rather than as negatives. The reported TVR improvements are meaningful, and the ablation study in Table 5 supports the contribution of each component. The paper also provides distribution analyses and qualitative examples. However, the central mechanism is not directly validated: the paper does not measure whether the detected ambiguous pairs are actually relevant to the query, and the ActivityNet gains are small and inconsistent across recall metrics. The final retrieval evaluation is external, so the reported numbers are not circular, but the proposed explanation for the gains is underdetermined. With additional validation experiments, this could be a solid contribution; in its current form, the main claim rests on an unverified assumption.

major comments (4)
  1. [Method, Eqs. (6)–(7); Experiments, Qualitative Results] The central claim that the detected ambiguous sets are genuinely relevant, rather than hard negatives, is not validated. Equations (6)–(7) define ambiguous pairs using the model's own similarity and uncertainty estimates, and the evidence for their semantic validity is limited to Fig. 4 (distributions), Fig. 5 (two qualitative examples), and Table 6 (an ablation over how to consume the detected set). None of these measures whether the detected unpaired videos contain a segment that matches the query. Because detection and training use the same online embeddings, the method could improve SumR simply by relaxing high-similarity negatives, without any semantic ambiguity being involved. A concrete fix is to use the moment/segment annotations available in TVR and ActivityNet Captions: for each detected ambiguous pair, check whether the video contains a segment matching the query, and report precision/recall of detection over training epochs. This would directly test the load-bearing assumption.
  2. [Tables 1 and 2] The statement that the method 'outperforms previous works in all recall metrics' is not supported by Table 1. On ActivityNet Captions with ResNet/I3D/RoBERTa features, the proposed method achieves R@5 = 24.6, which is lower than GMMFormer's 24.9, and R@1 is tied at 8.3. The SumR improvement is 2.3, but no error bars, multiple seeds, or significance tests are reported. Given that the ActivityNet gains are marginal and inconsistent across metrics, the claim of state-of-the-art performance on both datasets requires repeated runs and statistical testing.
  3. [Implementation Details] The thresholds τ_s and τ_u that define ambiguous pairs are set per epoch using statistics of the training set (mean of the positive-pair similarity distribution and mean of the uncertainty distribution). The paper provides no sensitivity analysis for these thresholds. Since the entire method depends on the detected ambiguous sets, the results could be sensitive to the choice of threshold. The authors should report performance as a function of τ_s and τ_u, or use a validation-based selection procedure, to demonstrate that the reported gains are not brittle.
  4. [Cross-Model Ambiguity Detection] The claim that reciprocal pseudo-label exchange between two identical encoders 'mitigates error propagation' is not directly tested. Table 5 rows (d) and (e) only compare the full model with and without cross-model detection, which shows a SumR improvement but does not isolate error-propagation mitigation from ensemble or regularization effects. A more direct test would compare single-model self-training with the same architecture against the proposed cross-model variant, and track the precision of detected ambiguous pairs over epochs to see whether error accumulation is actually reduced.
minor comments (6)
  1. [Abstract and Contributions] The phrase 'we achieves state-of-the-art performance' should be corrected to 'we achieve state-of-the-art performance'.
  2. [Eq. (3)] The word 'repec- tively' is a typo and should be 'respectively'.
  3. [Eqs. (8)–(10)] The set notation in the denominators, e.g., 'V ∈ A^q_i ∨ N^q_i', is ambiguous; it should be written as a union of sets, e.g., 'V ∈ A^q_i ∪ N^q_i', with explicit definitions of how negatives are sampled.
  4. [Fig. 3] The figure caption and diagram use several arrows and 'Logical AND' symbols without explaining the data flow clearly; please label the components or refer to specific equations in the caption.
  5. [Table 6] The column header 'A' is undefined; it should indicate the way ambiguous sets are used (e.g., 'Positive', 'Ignore', 'Ours').
  6. [Appendices] The text states 'More details are provided in the appendix,' but the appendix is not included in the manuscript; please include it or remove the reference.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the retrieval claims are evaluated against external ground truth, and the self-referential ambiguity detection is a training mechanism, not a derivation whose output equals its input.

full rationale

The paper's central claim is empirical: ARL improves partially relevant video retrieval on TVR and ActivityNet Captions, measured by rank-based recall against the datasets' ground-truth text-video pairs. The ambiguity detection in Eqs. 6-7 indeed uses the model's own similarity and uncertainty estimates, creating a self-referential pseudo-labeling loop, but this does not reduce the reported results to the method's inputs. The final retrieval scores (Eq. 13) are ranked against external annotations, and Table 6 shows that alternative treatments of the detected ambiguous sets (treating them all as positives, or ignoring them) perform worse, so the observed gains are not forced by construction. The paper also explicitly acknowledges the risk of error propagation when a model relies on its own predictions and proposes cross-model ambiguity detection as a mitigation. No load-bearing self-citations or imported uniqueness theorems appear; the related-work references to co-teaching and noisy-label learning are standard external sources. The main weakness is a validation gap: the paper does not quantitatively verify beyond two qualitative examples that detected ambiguous pairs are genuinely relevant, but that is a correctness or evidence concern, not a circularity of the kind defined by the analysis rules.

Assumptions & free parameters 5 free parameters · 4 assumptions · 0 invented entities

The framework depends on data-dependent thresholds, a small set of loss hyperparameters, and the domain assumption that high similarity plus high uncertainty indicates label ambiguity rather than a hard negative. No new physical or conceptual entities are introduced beyond the 'ambiguous pair' label, which is a training construct rather than an entity with independent evidence.

free parameters (5)
  • tau_s (similarity threshold) = Set to the mean of the positive-pair similarity distribution each epoch.
    Used in Eq 6-7 to flag ambiguous pairs; the value is fitted to the training set distribution.
  • tau_u (uncertainty threshold) = Set to the mean of the uncertainty distribution each epoch.
    Used with tau_s to detect ambiguous pairs; fitted to the training set distribution.
  • margin m_a = Not specified in the preprint; deferred to the appendix.
    Margin for the ambiguous triplet loss (Eq 11); must be smaller than m.
  • margin m = Not specified in the preprint; deferred to the appendix.
    Margin for the negative triplet loss (Eq 12).
  • lambda_nce = Not specified in the preprint.
    Balances the contrastive loss and triplet losses in Eq 10-12.
assumptions (4)
  • domain assumption Unpaired text-video pairs with high uncertainty and high similarity are ambiguous, not hard negatives.
    Eq 6-7 define ambiguity this way; the entire method rests on this premise.
  • domain assumption Uncertainty (average similarity across the dataset) measures how commonly shared an instance's context is.
    Eq 4 defines uncertainty; used to detect common context.
  • standard math Maximum frame similarity is a valid text-video retrieval score for partial relevance.
    Eq 2-3, from prior PRVR work.
  • ad hoc to paper Reciprocal pseudo-label exchange between two identical encoders reduces error propagation in self-training.
    Presented in Cross-Model Ambiguity Detection; assumed but not proven.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Ambiguity-Restrained Text-Video Representation Learning for Partially Relevant Video Retrieval." pith.science (2026). https://pith.science/paper/WULKKNXC

@misc{pith2026250607471,
  author       = {Pith},
  title        = {Pith review of: Ambiguity-Restrained Text-Video Representation Learning for Partially Relevant Video Retrieval},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/WULKKNXC}},
  note         = {Machine review of arXiv:2506.07471}
}
read the original abstract

Partially Relevant Video Retrieval~(PRVR) aims to retrieve a video where a specific segment is relevant to a given text query. Typical training processes of PRVR assume a one-to-one relationship where each text query is relevant to only one video. However, we point out the inherent ambiguity between text and video content based on their conceptual scope and propose a framework that incorporates this ambiguity into the model learning process. Specifically, we propose Ambiguity-Restrained representation Learning~(ARL) to address ambiguous text-video pairs. Initially, ARL detects ambiguous pairs based on two criteria: uncertainty and similarity. Uncertainty represents whether instances include commonly shared context across the dataset, while similarity indicates pair-wise semantic overlap. Then, with the detected ambiguous pairs, our ARL hierarchically learns the semantic relationship via multi-positive contrastive learning and dual triplet margin loss. Additionally, we delve into fine-grained relationships within the video instances. Unlike typical training at the text-video level, where pairwise information is provided, we address the inherent ambiguity within frames of the same untrimmed video, which often contains multiple contexts. This allows us to further enhance learning at the text-frame level. Lastly, we propose cross-model ambiguity detection to mitigate the error propagation that occurs when a single model is employed to detect ambiguous pairs for its training. With all components combined, our proposed method demonstrates its effectiveness in PRVR.

Figures

Figures reproduced from arXiv: 2506.07471 by the authors.

Figure 1
Figure 1. The illustration of ambiguous relationships be [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Illustration for ambiguous pair detection. To iden [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. Overview of ARL. (Left) Given the text-video train set, we initially calculate the text-video similarity to compute the uncertainty over the dataset level. Batch-wisely indexed uncertainty is used along the batch-wise similarity between texts and videos to explore ambiguous text-video relationships in the mini-batch. (Right) Two levels of Label Ambiguity Detection (LAD) modules detect ambiguous relationships. Text-V… view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Similarity and uncertainty distributions for positively-paired and unpaired text-video pairs on TVR. Similarity dis [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]
Figure 5
Figure 5. Figure 5: Ambiguity detection results on TVR dataset. For given queries, we visualize untrimmed videos that are detected to be [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

39 extracted references · 25 canonical work pages

  1. [1]

    , " * write output.state after.block = add.period write newline

    ENTRY address archivePrefix author booktitle chapter edition editor eid eprint howpublished institution isbn journal key month note number organization pages publisher school series title type volume year label extra.label sort.label short.list INTEGERS output.state before.all mid.sentence after.sentence after.block FUNCTION init.state.consts #0 'before.a...

  2. [2]

    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 global.max substring 't := if while FUNCTION word.in bbl.in capitalize " " * FUNCT...

  3. [3]

    Azadi, S.; Feng, J.; Jegelka, S.; and Darrell, T. 2015. Auxiliary image regularization for deep cnns with noisy labels. arXiv preprint arXiv:1511.07069

  4. [4]

    Balcan, M.-F.; Beygelzimer, A.; and Langford, J. 2006. Agnostic active learning. In Proceedings of the 23rd international conference on Machine learning, 65--72

  5. [5]

    Carreira, J.; and Zisserman, A. 2017. Quo vadis, action recognition? a new model and the kinetics dataset. In proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 6299--6308

  6. [6]

    Chen, T.; Kornblith, S.; Norouzi, M.; and Hinton, G. 2020. A simple framework for contrastive learning of visual representations. In International conference on machine learning, 1597--1607. PMLR

  7. [7]

    Dong, J.; Chen, X.; Zhang, M.; Yang, X.; Chen, S.; Li, X.; and Wang, X. 2022. Partially Relevant Video Retrieval. In Proceedings of the 30th ACM International Conference on Multimedia, 246--257

  8. [8]

    Dong, J.; Zhang, M.; Zhang, Z.; Chen, X.; Liu, D.; Qu, X.; Wang, X.; and Liu, B. 2023. Dual Learning with Dynamic Knowledge Distillation for Partially Relevant Video Retrieval. In Proceedings of the IEEE/CVF International Conference on Computer Vision, 11302--11312

Show all 39 references
  1. [9]

    J.; Kiros, J

    Faghri, F.; Fleet, D. J.; Kiros, J. R.; and Fidler, S. 2018. VSE++: Improving Visual-Semantic Embeddings with Hard Negatives

  2. [10]

    Fang, B.; Liu, C.; Zhou, Y.; Yang, M.; Song, Y.; Li, F.; Wang, W.; Ji, X.; Ouyang, W.; et al. 2023. Uatvr: Uncertainty-adaptive text-video retrieval. arXiv preprint arXiv:2301.06309

  3. [11]

    Freund, Y.; Schapire, R.; and Abe, N. 1999. A short introduction to boosting. Journal-Japanese Society For Artificial Intelligence, 14(771-780): 1612

  4. [12]

    K.; Vouitsis, N.; Ma, J.; Golestan, K.; Volkovs, M.; Garg, A.; and Yu, G

    Gorti, S. K.; Vouitsis, N.; Ma, J.; Golestan, K.; Volkovs, M.; Garg, A.; and Yu, G. 2022. X-pool: Cross-modal language-video attention for text-video retrieval. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 5006--5015

  5. [13]

    Han, B.; Yao, Q.; Yu, X.; Niu, G.; Xu, M.; Hu, W.; Tsang, I.; and Sugiyama, M. 2018. Co-teaching: Robust training of deep neural networks with extremely noisy labels. Advances in neural information processing systems, 31

  6. [14]

    He, K.; Zhang, X.; Ren, S.; and Sun, J. 2016. Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition, 770--778

  7. [15]

    Hou, Z.; Ngo, C.-W.; and Chan, W. K. 2021. CONQUER: Contextual query-aware ranking for video corpus moment retrieval. In Proceedings of the 29th ACM International Conference on Multimedia, 3900--3908

  8. [16]

    Jin, P.; Li, H.; Cheng, Z.; Huang, J.; Wang, Z.; Yuan, L.; Liu, C.; and Chen, J. 2023. Text-video retrieval with disentangled conceptualization and set-to-set alignment. arXiv preprint arXiv:2305.12218

  9. [17]

    Ju, L.; Wang, X.; Wang, L.; Mahapatra, D.; Zhao, X.; Zhou, Q.; Liu, T.; and Ge, Z. 2022. Improving medical images classification with label noise using dual-uncertainty estimation. IEEE transactions on medical imaging, 41(6): 1533--1546

  10. [18]

    Khosla, P.; Teterwak, P.; Wang, C.; Sarna, A.; Tian, Y.; Isola, P.; Maschinot, A.; Liu, C.; and Krishnan, D. 2020. Supervised contrastive learning. Advances in neural information processing systems, 33: 18661--18673

  11. [19]

    Krishna, R.; Hata, K.; Ren, F.; Fei-Fei, L.; and Carlos Niebles, J. 2017. Dense-captioning events in videos. In Proceedings of the IEEE international conference on computer vision, 706--715

  12. [20]

    L.; and Bansal, M

    Lei, J.; Yu, L.; Berg, T. L.; and Bansal, M. 2020. Tvr: A large-scale dataset for video-subtitle moment retrieval. In Computer Vision--ECCV 2020: 16th European Conference, Glasgow, UK, August 23--28, 2020, Proceedings, Part XXI 16, 447--463. Springer

  13. [21]

    Li, H.; Song, J.; Gao, L.; Zhu, X.; and Shen, H. 2024. Prototype-based Aleatoric Uncertainty Quantification for Cross-modal Retrieval. Advances in Neural Information Processing Systems, 36

  14. [22]

    Li, J.; Socher, R.; and Hoi, S. C. 2020. Dividemix: Learning with noisy labels as semi-supervised learning. arXiv preprint arXiv:2002.07394

  15. [23]

    Lin, C.; Wu, A.; Liang, J.; Zhang, J.; Ge, W.; Zheng, W.-S.; and Shen, C. 2022. Text-adaptive multiple visual prototype matching for video-text retrieval. Advances in neural information processing systems, 35: 38655--38666

  16. [24]

    Lin, Y.; Zhang, J.; Huang, Z.; Liu, J.; Wen, Z.; and Peng, X. 2024. Multi-granularity correspondence learning from long-term noisy videos. arXiv preprint arXiv:2401.16702

  17. [25]

    Liu, Y.; Ott, M.; Goyal, N.; Du, J.; Joshi, M.; Chen, D.; Levy, O.; Lewis, M.; Zettlemoyer, L.; and Stoyanov, V. 2019. Roberta: A robustly optimized bert pretraining approach. arXiv preprint arXiv:1907.11692

  18. [26]

    Ma, Y.; Xu, G.; Sun, X.; Yan, M.; Zhang, J.; and Ji, R. 2022. X-clip: End-to-end multi-grained contrastive learning for video-text retrieval. In Proceedings of the 30th ACM International Conference on Multimedia, 638--647

  19. [27]

    Neverova, N.; Novotny, D.; and Vedaldi, A. 2019. Correlated uncertainty for learning dense correspondences from noisy labels. Advances in Neural Information Processing Systems, 32

  20. [28]

    Northcutt, C.; Jiang, L.; and Chuang, I. 2021. Confident learning: Estimating uncertainty in dataset labels. Journal of Artificial Intelligence Research, 70: 1373--1411

  21. [29]

    W.; Hallacy, C.; Ramesh, A.; Goh, G.; Agarwal, S.; Sastry, G.; Askell, A.; Mishkin, P.; Clark, J.; et al

    Radford, A.; Kim, J. W.; Hallacy, C.; Ramesh, A.; Goh, G.; Agarwal, S.; Sastry, G.; Askell, A.; Mishkin, P.; Clark, J.; et al. 2021. Learning transferable visual models from natural language supervision. In International conference on machine learning, 8748--8763. PMLR

  22. [30]

    Song, Y.; and Soleymani, M. 2019. Polysemous visual-semantic embedding for cross-modal retrieval. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 1979--1988

  23. [31]

    Tan, C.; Xia, J.; Wu, L.; and Li, S. Z. 2021. Co-learning: Learning from noisy labels with self-supervision. In Proceedings of the 29th ACM International Conference on Multimedia, 1405--1413

  24. [32]

    Wang, Q.; Zhang, Y.; Zheng, Y.; Pan, P.; and Hua, X.-S. 2022. Disentangled Representation Learning for Text-Video Retrieval. arXiv:2203.07111

  25. [33]

    Wang, Y.; Ma, X.; Chen, Z.; Luo, Y.; Yi, J.; and Bailey, J. 2019. Symmetric cross entropy for robust learning with noisy labels. In Proceedings of the IEEE/CVF international conference on computer vision, 322--330

  26. [34]

    Wang, Y.; Wang, J.; Chen, B.; Zeng, Z.; and Xia, S.-T. 2024. GMMFormer: Gaussian-Mixture-Model based Transformer for Efficient Partially Relevant Video Retrieval. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 38, 5767--5775

  27. [35]

    Wei, H.; Feng, L.; Chen, X.; and An, B. 2020. Combating noisy labels by agreement: A joint training method with co-regularization. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 13726--13735

  28. [36]

    Wu, W.; Luo, H.; Fang, B.; Wang, J.; and Ouyang, W. 2023. Cap4video: What can auxiliary captions do for text-video retrieval? In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 10704--10713

  29. [37]

    Zhang, B.; Hu, H.; Lee, J.; Zhao, M.; Chammas, S.; Jain, V.; Ie, E.; and Sha, F. 2020. A hierarchical multi-modal encoder for moment localization in video corpus. arXiv preprint arXiv:2011.09046

  30. [38]

    T.; and Goh, R

    Zhang, H.; Sun, A.; Jing, W.; Nan, G.; Zhen, L.; Zhou, J. T.; and Goh, R. S. M. 2021. Video corpus moment retrieval with contrastive learning. In Proceedings of the 44th International ACM SIGIR Conference on Research and Development in Information Retrieval, 685--695

  31. [39]

    Zheng, Z.; and Yang, Y. 2021. Rectifying pseudo label learning via uncertainty estimation for domain adaptive semantic segmentation. International Journal of Computer Vision, 129(4): 1106--1120

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.