Pith. sign in

REVIEW 3 major objections 5 minor 55 references

Dynamic Adapter with Semantics Disentangling for Cross-lingual Cross-modal Retrieval

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 DASD generates adapter weights from each target-language caption's disentangled semantics and style, and the paper reports state-of-the-art cross-lingual image-text and video-text retrieval results with only 134M trainable parameters.

desk verdict Useful empirical paper on dynamic adapters for cross-lingual retrieval, but its central disentangling loss does not enforce what it claims, so the mechanism is oversold. read the letter →

arxiv 2412.13510 v1 pith:OY6IOM6V submitted 2024-12-18 cs.CV cs.CL

Rui Cai , Zhiyu Dong , Jianfeng Dong , Xun Wang This is my paper · ORCID
classification cs.CVcs.CL
keywords cross-lingualcross-modalretrievaldynamicadaptersemanticsdisentanglingparameter-efficientfine-tuningvision-languagepretrainingimage-textvideo-textlow-resourcelanguages
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 tries to show that cross-lingual cross-modal retrieval can be made both parameter-efficient and adaptive by generating adapter weights fresh for every input caption. It proposes DASD, which splits each target-language caption into semantic-related and semantic-agnostic features using a small disentangling module built on one frozen CLIP layer, then uses the concatenation of those features to synthesize adapter parameters inserted into a frozen text encoder. The authors report that DASD outperforms static adapter methods such as MLA and full-model fine-tuning methods such as NRCCR, DCOT, and CL2CM on Multi30K, MSCOCO, and MSRVTT, across eight target languages, under both cross-lingual fine-tuning and zero-shot settings. If true, this provides a low-resource route to extending frozen vision-language models to new languages without large labeled target-language data.

What carries the argument

The central object is the input-aware dynamic adapter paired with a semantics disentangling module. The semantics disentangling module uses only the first frozen CLIP text layer plus two trainable bottleneck MLPs, the semantic-related adapter $A_{sr}$ and the semantic-agnostic adapter $A_{sa}$, to produce $f^{sr}$ and $f^{sa}$; an adversarial classifier tries to distinguish $(f^{sa}, r_S)$ from $(f^{sa}, r_{S^-})$, forcing $f^{sa}$ to shed semantic content, while an L1 loss pushes $f^{sr}$ close to the source representation $r_S$. These features are concatenated into $z$, and a linear down-projection $W^{down}_i$ reshapes a vector into the parameter matrix $W^z_i$ of the bottleneck dynamic adapter $DA(X)=W^{upper}_d\,\mathrm{ReLU}(W^z W^{down}_d X)$. The dynamic adapter is inserted at every layer of the frozen target-language text encoder, so each caption receives its own adapter weights while the pretrained model stays frozen.

What would settle it

Train a classifier on the extracted semantic-agnostic features to predict the caption's semantic content, such as its source-language representation or image category: if accuracy stays high while the paper's losses are active, the adversarial disentangling has not blocked semantic leakage, and the dynamic adapter's gains should be re-attributed.

Watch

Extended reading notes

Core claim

The central claim is that the right unit of cross-lingual transfer is not a language-specific static adapter but an input-conditioned adapter whose weights are generated from the caption itself. DASD defines two complementary views of a target-language caption: a semantic-related feature $f^{sr}$ aligned to source-language semantics through L1 consistency, and a semantic-agnostic feature $f^{sa}$ trained adversarially to be uninformative about semantics yet to capture word order and expression. The two features are concatenated and mapped through an MLP and a low-rank down-projection to produce the parameter matrix $W^z_i$ of a dynamic adapter at each transformer layer of the frozen CLIP text encoder. Training combines cross-lingual MSE alignment, cross-modal NCE contrastive loss, the adversarial loss, and the semantic consistency loss. The paper reports state-of-the-art mAR scores on Multi30K and MSCOCO, text-to-video Recall@1 on MSRVTT, and improved results when the dynamic adapter is attached to CCLM rather than CLIP, with 134M trainable parameters compared with 108M for MLA.

Load-bearing premise

The load-bearing premise is that the small disentangling module really separates a caption's meaning from its way of saying it, so the generated adapter weights match the caption's true characteristics.

Editorial extensions

If this is right

  • DASD's dynamic adapters beat the static MLA adapter on all eight target languages tested for video-text retrieval, under both zero-shot and cross-lingual fine-tuning.
  • The method stays parameter-efficient: 134M trainable parameters compared with 108M for MLA and far fewer than full-model methods, while outperforming full-model baselines such as NRCCR, DCOT, and CL2CM.
  • Swapping the frozen backbone from CLIP to the stronger CCLM raises the summed image-text retrieval score from 446.3 to 452.7, so the dynamic adapters transfer across VLP families.
  • Dynamic parameters themselves carry the main gain: replacing them with a static adapter drops summed mAR from 446.3 to 429.4.
  • Both disentangled features matter: removing either $f^{sr}$ or $f^{sa}$ lowers summed mAR by roughly 4 to 5 points, and ablating the two disentangling losses also lowers it.

Reading between the lines

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

  • If the disentangling genuinely captures expression style, the same parameter-generation mechanism could be applied to adapters in machine translation, style transfer, or domain generalization, where input-dependent shifts are the main obstacle.
  • A testable extension is to use the semantic-agnostic feature to deliberately control expression: editing $f^{sa}$ while keeping $f^{sr}$ fixed should change word order and length without changing the caption's meaning.
  • The reported gains over static adapters may partly reflect added capacity and input-dependence rather than true semantic separation; an ablation that randomizes the mapping from features to adapter weights while keeping capacity fixed would separate those effects.
  • Because the method only needs the first frozen layer of the text encoder plus two small MLPs, it should extend naturally to other low-resource languages beyond the eight evaluated, as long as a multilingual tokenizer covers them.
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 DASD, a dynamic adapter framework for cross-lingual cross-modal retrieval (CCR). The method freezes a CLIP text encoder, inserts lightweight adapters into each layer of the target-language branch, and generates the adapter parameters per input caption. A semantics disentangling module, built from the first frozen CLIP layer plus two trainable adapters, produces a semantic-related feature f_sr and a semantic-agnostic feature f_sa via L1 consistency with the source representation and an adversarial loss. The concatenated features are used to generate a dynamic parameter matrix for the adapters. Training combines cross-lingual alignment (MSE between source and target text representations), cross-modal alignment (NCE loss), and the two disentangling losses. Experiments on Multi30K, MSCOCO, and MSRVTT report consistent gains over static adapters and prior full-model methods across several target languages, plus compatibility with the CCLM VLP model.

Significance. If the central claim holds, DASD is a practically useful parameter-efficient way to extend frozen vision-language pretrained models to new languages, addressing an important problem. The paper is clearly written and specifies the method precisely, with loss formulas, ablations for each component (Tables 4-6), and a supplementary study of design choices. The authors also release code, which aids reproducibility. The main weakness is that the key disentangling mechanism is not convincingly validated: the adversarial loss in Eq. (7) does not necessarily enforce a semantic-agnostic f_sa, and the ablations only show that f_sa is useful, not that it is invariant to semantics. The empirical evaluation also lacks error bars and appears to select hyperparameters on the test sets, which weakens the strength of the SOTA claims.

major comments (3)
  1. The adversarial loss in Eq. (7) does not enforce that f_sa is semantic-agnostic. The discriminator F is trained on pairs (f_sa, rS) and (f_sa, rS-), where f_sa is identical in both pairs. A sufficiently expressive discriminator can achieve near-perfect accuracy by ignoring f_sa and simply distinguishing rS from rS-; once this happens, the adversarial gradient w.r.t. f_sa vanishes, so no pressure removes semantic information from f_sa. The paper's statement that the goal is 'perfect semantics disentangling' is therefore unsupported. The ablation in Table 5 shows only that both f_sr and f_sa contribute to retrieval performance, which would also hold if f_sa leaked semantic content. The t-SNE visualization in Figure 3 is qualitative and does not control for topic confounds. Please provide a direct invariance test, e.g., train a classifier on f_sa to predict the associated image or source caption and show chance-level accuracy, or use a mutual-information estimator against semantic labels, and report the result.
  2. Hyperparameters appear to be selected using test-set performance. Table 7 compares L1, L2, and Smooth-L1 losses and Table 11 varies the dynamic adapter dimension du, reporting mAR on the Multi30K and MSCOCO test sets without a validation split; the main-text choices λ1=1 and λ2=0.1 are also presented without a selection procedure. This practice can overfit the test sets and makes the differences in Table 4 (dynamic vs. static adapter) difficult to interpret, since the static baseline is not given equivalent test-set tuning. Please perform model selection on a validation split and report mean and standard deviation over at least three random seeds for the main results.
  3. The reported improvements are not accompanied by error bars or significance tests. Some gains are substantial (e.g., Table 2, SW: 22.1 vs. 12.6 in CL-FT), but single-run results cannot establish that differences are not due to noise. This is especially relevant for the core comparison in Table 4, where the dynamic adapter improvement over the traditional adapter varies by language from 2.8 to 4.8 points. Please report multiple seeds with standard deviations, or at least confirm the main conclusions with a paired significance test.
minor comments (5)
  1. Equation (7) is missing a closing parenthesis: it should be -log(1 - F(f_sa, rS-)).
  2. The phrase 'Considering the the scarcity' contains a duplicated 'the'. Please correct the typo.
  3. The full-model sum in Table 10 is 446.8, while the same configuration in Table 4 is 446.3. Please reconcile the numbers.
  4. In the zero-shot block, DASD with MT(MSCOCO) training reports no entries for ZH and JA. The table uses dashes for these languages across methods in that row, but a brief note explaining that MT(MSCOCO) does not include these target languages would help the reader.
  5. The paper states the code is 'anonymously released' but also gives a non-anonymous GitHub URL (https://github.com/HuiGuanLab/DASD). Please clarify the anonymity status or remove the word 'anonymously'.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: DASD's claims are empirical and its components are evaluated by controlled ablations and external benchmarks, not by reusing fitted values or self-citations as evidence.

full rationale

The paper's central claims are that dynamic, input-conditioned adapters improve cross-lingual cross-modal retrieval over static adapters, and that semantic-related and semantic-agnostic features with disentangling losses drive the gain. Neither claim reduces to its inputs by construction. The dynamic parameter matrix is generated from f_sr and f_sa via Equations 8-9, but the model is trained with cross-lingual MSE, contrastive NCE, and disentangling losses (Equations 14-16) and then evaluated on held-out retrieval benchmarks in multiple target languages; success is not defined by the fitted losses themselves. The ablations in Tables 4-6 remove W_z, f_sa/f_sr, and Ladv/Lsc, and measure retrieval mAR, so the contribution of each component is an empirical finding rather than a renamed input. Self-citations to the authors' NRCCR, DCOT, and CL2CM lines appear, but these are used as comparison baselines rather than as load-bearing evidence for DASD's design; no central argument is outsourced to a self-citation. The adversarial loss in Equation 7 may be open to a validity objection, since a discriminator could in principle rely on rS versus rS- alone without reading f_sa, but that is a correctness or robustness concern about whether disentangling is achieved, not a circular step in which a prediction is equivalent to its input. The paper is self-contained against external benchmarks, so no circularity score above 0 is warranted.

Assumptions & free parameters 5 free parameters · 4 assumptions · 2 invented entities

The method introduces no new physical entities; the free parameters are standard ML hyperparameters, some tuned on test-set ablations. The key postulates are the separability of semantic and style features using the proposed losses and the sufficiency of a single frozen layer for disentangling, both supported only by internal experiments. The reliance on MT-generated data is an external dataset assumption shared by the CCR setting.

free parameters (5)
  • lambda1 (adversarial loss weight) = 1
    Chosen by hand; no ablation reported for these weights.
  • lambda2 (semantic consistency loss weight) = 0.1
    Chosen by hand; no ablation reported.
  • d_u (dynamic adapter hidden dimension) = 32
    Selected by test-set ablation in Table 11; 16 is worse, 64 gives negligible gain.
  • SDM backbone layers = 1
    Selected by ablation (Figure 4); 0 layers drops performance sharply, more layers give no clear gain.
  • temperature tau in NCE loss = 0.01
    Set following common practice; no sensitivity study.
assumptions (4)
  • domain assumption Frozen CLIP text and image encoders provide a sufficiently aligned semantic space for source-language captions.
    The entire CCR setup relies on the VLP model having learned English vision-language alignment; used throughout the Proposed Method.
  • domain assumption Machine-translated target-language captions are a reliable proxy for human-written target captions.
    All target-language training data is obtained via Google Translate (Datasets section); if MT noise is severe, the alignments learned are degraded.
  • ad hoc to paper Semantic-related and semantic-agnostic features are independent and can be separated by L1 consistency plus adversarial training.
    The core design premise of the disentangling module (Eqs. 5-7); no theoretical guarantee is given.
  • ad hoc to paper A single frozen transformer layer is sufficient to extract features for adapter generation.
    Supported only by the paper's own ablation (Figure 4), not by an independent principle.
invented entities (2)
  • Semantic-related feature f_sr
    purpose: Caption representation that carries the meaning, used to generate adapter parameters.
    Defined operationally by Eq. 5 and L1 loss; validated only by internal t-SNE and ablations.
  • Semantic-agnostic feature f_sa
    purpose: Caption representation capturing expression style (word order, length) used to generate adapter parameters.
    Defined by average pooling and adversarial training (Eq. 7); the paper's own Figure 3 clusters are qualitative and computed on the test set.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Dynamic Adapter with Semantics Disentangling for Cross-lingual Cross-modal Retrieval." pith.science (2026). https://pith.science/paper/OY6IOM6V

@misc{pith2026241213510,
  author       = {Pith},
  title        = {Pith review of: Dynamic Adapter with Semantics Disentangling for Cross-lingual Cross-modal Retrieval},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/OY6IOM6V}},
  note         = {Machine review of arXiv:2412.13510}
}
read the original abstract

Existing cross-modal retrieval methods typically rely on large-scale vision-language pair data. This makes it challenging to efficiently develop a cross-modal retrieval model for under-resourced languages of interest. Therefore, Cross-lingual Cross-modal Retrieval (CCR), which aims to align vision and the low-resource language (the target language) without using any human-labeled target-language data, has gained increasing attention. As a general parameter-efficient way, a common solution is to utilize adapter modules to transfer the vision-language alignment ability of Vision-Language Pretraining (VLP) models from a source language to a target language. However, these adapters are usually static once learned, making it difficult to adapt to target-language captions with varied expressions. To alleviate it, we propose Dynamic Adapter with Semantics Disentangling (DASD), whose parameters are dynamically generated conditioned on the characteristics of the input captions. Considering that the semantics and expression styles of the input caption largely influence how to encode it, we propose a semantic disentangling module to extract the semantic-related and semantic-agnostic features from the input, ensuring that generated adapters are well-suited to the characteristics of input caption. Extensive experiments on two image-text datasets and one video-text dataset demonstrate the effectiveness of our model for cross-lingual cross-modal retrieval, as well as its good compatibility with various VLP models.

Figures

Figures reproduced from arXiv: 2412.13510 by the authors.

Figure 1
Figure 1. Illustration of the variety of textual expressions [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. The illustration of our proposed Dynamic Adapter with Semantics Disentangling (DASD). To make dynamic adapters [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Visualization of the semantic-agnostic features ex [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Performance of our model varies with the number [PITH_FULL_IMAGE:figures/full_fig_p012_4.png]
Figure 5
Figure 5. Figure 5: Visualization of the semantic-related features ex [PITH_FULL_IMAGE:figures/full_fig_p013_5.png]
Figure 6
Figure 6. Figure 6: Supplementary samples within the scope of group 1 and group 2 in the main text Figure 3(a). [PITH_FULL_IMAGE:figures/full_fig_p014_6.png]
Figure 7
Figure 7. Figure 7: Supplementary samples within the scope of group 3 and group 4 in the main text Figure 3(a). [PITH_FULL_IMAGE:figures/full_fig_p015_7.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

55 extracted references · 37 canonical work pages

  1. [1]

    Aggarwal, P.; and Kale, A. 2020. Towards zero-shot cross-lingual image retrieval. arXiv preprint arXiv:2012.05107

    arXiv 2020

  2. [2]

    U.; Zhang, Z.; Ma, X.; Chang, K.-W.; and Peng, N

    Ahmad, W. U.; Zhang, Z.; Ma, X.; Chang, K.-W.; and Peng, N. 2019. Cross-Lingual Dependency Parsing with Unlabeled Auxiliary Languages. In Bansal, M.; and Villavicencio, A., eds., Proceedings of the 23rd Conference on Computational Natural Language Learning (CoNLL), 372--382. Association for Computational Linguistics

    work page 2019

  3. [3]

    M.; Pfeiffer, J.; Ruder, S.; Glava s , G.; Vuli \'c , I.; and Korhonen, A

    Ansell, A.; Ponti, E. M.; Pfeiffer, J.; Ruder, S.; Glava s , G.; Vuli \'c , I.; and Korhonen, A. 2021. MAD-G: Multilingual adapter generation for efficient cross-lingual transfer. In Findings of the Association for Computational Linguistics: EMNLP 2021, 4762--4781

    work page 2021

  4. [4]

    Bogolin, S.-V.; Croitoru, I.; Jin, H.; Liu, Y.; and Albanie, S. 2022. Cross modal retrieval with querybank normalisation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 5194--5205

    work page 2022

  5. [5]

    Cai, R.; Dong, J.; Liang, T.; Liang, Y.; Wang, Y.; Yang, X.; Wang, X.; and Wang, M. 2024. Cross-Lingual Cross-Modal Retrieval with Noise-Robust Fine-Tuning. IEEE Transactions on Knowledge and Data Engineering

    work page 2024

  6. [6]

    Cai, R.; Li, Z.; Wei, P.; Qiao, J.; Zhang, K.; and Hao, Z. 2019. Learning disentangled semantic representation for domain adaptation. In IJCAI: proceedings of the conference, volume 2019, 2060. NIH Public Access

    work page 2019

  7. [7]

    Chang, T.; Yang, X.; Luo, X.; Ji, W.; and Wang, M. 2023. Learning Style-Invariant Robust Representation for Generalizable Visual Instance Retrieval. In Proceedings of the 31st ACM International Conference on Multimedia, 6171--6180

    work page 2023

  8. [8]

    Chen, X.; Fang, H.; Lin, T.-Y.; Vedantam, R.; Gupta, S.; Doll \'a r, P.; and Zitnick, C. L. 2015. Microsoft coco captions: Data collection and evaluation server. arXiv preprint arXiv:1504.00325

    arXiv 2015

Show all 55 references
  1. [9]

    Chen, Z.; Luo, Y.; Qiu, R.; Wang, S.; Huang, Z.; Li, J.; and Zhang, Z. 2021. Semantics disentangling for generalized zero-shot learning. In Proceedings of the IEEE/CVF international conference on computer vision, 8712--8720

    2021

  2. [10]

    Devlin, J.; Chang, M.-W.; Lee, K.; and Toutanova, K. 2018. Bert: Pre-training of deep bidirectional transformers for language understanding. arXiv preprint arXiv:1810.04805

    2018 arXiv

  3. [11]

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

    2022

  4. [12]

    Dong, J.; Li, X.; Xu, C.; Yang, X.; Yang, G.; Wang, X.; and Wang, M. 2022 b . Dual encoding for video retrieval by text. IEEE Transactions on Pattern Analysis and Machine Intelligence, 44(8): 4065--4080

    2022

  5. [13]

    Dosovitskiy, A.; Beyer, L.; Kolesnikov, A.; Weissenborn, D.; Zhai, X.; Unterthiner, T.; Dehghani, M.; Minderer, M.; Heigold, G.; Gelly, S.; et al. 2020. An image is worth 16x16 words: Transformers for image recognition at scale. International Conference on Leanring Representations

    2020

  6. [14]

    Elliott, D.; Frank, S.; Sima'an, K.; and Specia, L. 2016. Multi30k: Multilingual english-german image descriptions. In Proceedings of the 5th Workshop on Vision and Language, 70--74

    2016

  7. [15]

    Fang, X.; Easwaran, A.; Genest, B.; and Suganthan, P. N. 2024. Your Data Is Not Perfect: Towards Cross-Domain Out-of-Distribution Detection in Class-Imbalanced Data. ESWA

    2024

  8. [16]

    Fang, X.; Liu, D.; Zhou, P.; and Nan, G. 2023. You can ground earlier than see: An effective and efficient pipeline for temporal sentence grounding in compressed videos. In CVPR

    2023

  9. [17]

    Gutmann, M.; and Hyv \"a rinen, A. 2010. Noise-contrastive estimation: A new estimation principle for unnormalized statistical models. In Proceedings of the thirteenth international conference on artificial intelligence and statistics, 297--304. JMLR Workshop and Conference Pr...

    2010

  10. [18]

    Hedegaard, L.; Alok, A.; Jose, J.; and Iosifidis, A. 2024. Structured pruning adapters. Pattern Recognition, 156: 110724

    2024

  11. [19]

    Huang, P.-Y.; Patrick, M.; Hu, J.; Neubig, G.; Metze, F.; and Hauptmann, A. G. 2021. Multilingual Multimodal Pre-training for Zero-Shot Cross-Lingual Transfer of Vision-Language Models. In Proceedings of the 2021 Conference of the North American Chapter of the Association for ...

    2021

  12. [20]

    Jain, A.; Guo, M.; Srinivasan, K.; Chen, T.; Kudugunta, S.; Jia, C.; Yang, Y.; and Baldridge, J. 2021. MURAL: multimodal, multitask retrieval across languages. In Findings of the Association for Computational Linguistics: EMNLP 2021, 3449--3463

    2021

  13. [21]

    Li, H.; Zhang, C.; Jia, X.; Gao, Y.; and Chen, C. 2021. Adaptive label correlation based asymmetric discrete hashing for cross-modal retrieval. IEEE Transactions on Knowledge and Data Engineering, 35(2): 1185--1199

    2021

  14. [22]

    Li, X.; Xu, C.; Wang, X.; Lan, W.; Jia, Z.; Yang, G.; and Xu, J. 2019. COCO-CN for cross-lingual image tagging, captioning, and retrieval. IEEE Transactions on Multimedia, 21(9): 2347--2360

    2019

  15. [23]

    L.; and Liang, P

    Li, X. L.; and Liang, P. 2021. Prefix-tuning: Optimizing continuous prompts for generation. arXiv preprint arXiv:2101.00190

    2021 arXiv

  16. [24]

    Liu, X.; Zheng, Y.; Du, Z.; Ding, M.; Qian, Y.; Yang, Z.; and Tang, J. 2024. GPT understands, too. AI Open, 5: 208--215

    2024

  17. [25]

    Lu, H.; Fei, N.; Huo, Y.; Gao, Y.; Lu, Z.; and Wen, J.-R. 2022. Cots: Collaborative two-stream vision-language pre-training model for cross-modal retrieval. In Proceedings of the IEEE/CVF conference on computer Vision and pattern recognition, 15692--15701

    2022

  18. [26]

    Luo, H.; Ji, L.; Zhong, M.; Chen, Y.; Lei, W.; Duan, N.; and Li, T. 2022. Clip4clip: An empirical study of clip for end to end video clip retrieval and captioning. Neurocomputing, 508: 293--304

    2022

  19. [27]

    Mao, Y.; Huang, K.; Guan, C.; Bao, G.; Mo, F.; and Xu, J. 2024. DoRA: Enhancing Parameter-Efficient Fine-Tuning with Dynamic Rank Distribution. arXiv preprint arXiv:2405.17357

    2024 arXiv

  20. [28]

    Ni, M.; Huang, H.; Su, L.; Cui, E.; Bharti, T.; Wang, L.; Zhang, D.; and Duan, N. 2021. M3p: Learning universal representations via multitask multilingual multimodal pre-training. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 3977--3986

    2021

  21. [29]

    Pfeiffer, J.; Vuli \'c , I.; Gurevych, I.; and Ruder, S. 2020. MAD-X: An Adapter-Based Framework for Multi-Task Cross-Lingual Transfer. Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP), 7654--7673

    2020

  22. [30]

    Portaz, M.; Randrianarivo, H.; Nivaggioli, A.; Maudet, E.; Servan, C.; and Peyronnet, S. 2019. Image search using multilingual texts: a cross-modal learning approach between image and text. arXiv preprint arXiv:1903.11299

    2019 arXiv

  23. [31]

    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

    2021

  24. [32]

    Rouditchenko, A.; Chuang, Y.-S.; Shvetsova, N.; Thomas, S.; Feris, R.; Kingsbury, B.; Karlinsky, L.; Harwath, D.; Kuehne, H.; and Glass, J. 2023. C2kd: Cross-lingual cross-modal knowledge distillation for multilingual text-video retrieval. In ICASSP 2023-2023 IEEE Internationa...

    2023

  25. [33]

    Sharma, P.; Ding, N.; Goodman, S.; and Soricut, R. 2018. Conceptual captions: A cleaned, hypernymed, image alt-text dataset for automatic image captioning. In Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), 2556--2565

    2018

  26. [34]

    Sun, Y.; Ren, Z.; Hu, P.; Peng, D.; and Wang, X. 2023. Hierarchical consensus hashing for cross-modal retrieval. IEEE Transactions on Multimedia, 26: 824--836

    2023

  27. [35]

    N.; Kaiser, .; and Polosukhin, I

    Vaswani, A.; Shazeer, N.; Parmar, N.; Uszkoreit, J.; Jones, L.; Gomez, A. N.; Kaiser, .; and Polosukhin, I. 2017. Attention is all you need. Advances in neural information processing systems, 30

    2017

  28. [36]

    Wang, Y.; Dong, J.; Liang, T.; Zhang, M.; Cai, R.; and Wang, X. 2022. Cross-lingual cross-modal retrieval with noise-robust learning. In Proceedings of the 30th ACM International Conference on Multimedia, 422--433

    2022

  29. [37]

    Wang, Y.; Wang, F.; Dong, J.; and Luo, H. 2024 a . CL2CM: Improving Cross-Lingual Cross-Modal Retrieval via Cross-Lingual Knowledge Transfer. Proceedings of the AAAI Conference on Artificial Intelligence, 38(6): 5651--5659

    2024

  30. [38]

    Wang, Y.; Wang, S.; Luo, H.; Dong, J.; Wang, F.; Han, M.; Wang, X.; and Wang, M. 2024 b . Dual-view Curricular Optimal Transport for Cross-lingual Cross-modal Retrieval. IEEE Transactions on Image Processing, 33: 1522--1533

    2024

  31. [39]

    Wu, C.; Jiang, G.; and Lian, D. 2024. Mitigate Negative Transfer with Similarity Heuristic Lifelong Prompt Tuning. arXiv preprint arXiv:2406.12251

    2024 arXiv

  32. [40]

    Xu, J.; Mei, T.; Yao, T.; and Rui, Y. 2016. MSR-VTT : A Large Video Description Dataset for Bridging Video and Language. In Proceedings of the IEEE/CVF conference on Computer Vision and Pattern Recognition, 5288--5296

    2016

  33. [41]

    Ye, Z.; Hu, F.; Lyu, F.; Li, L.; and Huang, K. 2021. Disentangling semantic-to-visual confusion for zero-shot learning. IEEE Transactions on Multimedia, 24: 2828--2840

    2021

  34. [42]

    Yin, G.; Liu, B.; Sheng, L.; Yu, N.; Wang, X.; and Shao, J. 2019. Semantics disentangling for text-to-image generation. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2327--2336

    2019

  35. [43]

    Yoshikawa, Y.; Shigeto, Y.; and Takeuchi, A. 2017. STAIR captions: Constructing a large-scale Japanese image caption dataset. arXiv preprint arXiv:1705.00823

    2017 arXiv

  36. [44]

    Young, P.; Lai, A.; Hodosh, M.; and Hockenmaier, J. 2014. From image descriptions to visual denotations: New similarity metrics for semantic inference over event descriptions. Transactions of the Association for Computational Linguistics, 2: 67--78

    2014

  37. [45]

    Zeng, Y.; Zhou, W.; Luo, A.; Cheng, Z.; and Zhang, X. 2023. Cross-View Language Modeling: Towards Unified Cross-Lingual Cross-Modal Pre-training. In Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), 5731--5746

    2023

  38. [46]

    Zhang, A.; Wang, H.; Wang, X.; and Chua, T.-S. 2024. Disentangling Masked Autoencoders for Unsupervised Domain Generalization. arXiv preprint arXiv:2407.07544

    2024 arXiv

  39. [47]

    Zhang, D.; Zhang, H.; Tang, J.; Hua, X.-S.; and Sun, Q. 2020 a . Causal intervention for weakly-supervised semantic segmentation. Advances in Neural Information Processing Systems, 33: 655--666

    2020

  40. [48]

    Zhang, D.; Zhang, H.; Tang, J.; Wang, M.; Hua, X.; and Sun, Q. 2020 b . Feature pyramid transformer. In Computer Vision--ECCV 2020: 16th European Conference, Glasgow, UK, August 23--28, 2020, Proceedings, Part XXVIII 16, 323--339. Springer

    2020

  41. [49]

    Zhang, L.; Hu, A.; and Jin, Q. 2022. Multi-Lingual Acquisition on Multimodal Pre-training for Cross-modal Retrieval. Advances in Neural Information Processing Systems, 35: 29691--29704

    2022

  42. [50]

    Zhang, Z.; Luo, H.; Zhu, L.; Lu, G.; and Shen, H. T. 2023. Modality-Invariant Asymmetric Networks for Cross-Modal Hashing. IEEE Transactions on Knowledge and Data Engineering, 35(5): 5091--5104

    2023

  43. [51]

    Zheng, Q.; Dong, J.; Qu, X.; Yang, X.; Wang, Y.; Zhou, P.; Liu, B.; and Wang, X. 2023. Progressive localization networks for language-based moment localization. ACM Transactions on Multimedia Computing, Communications and Applications, 19(2): 1--21

    2023

  44. [52]

    C.; and Liu, Z

    Zhou, K.; Yang, J.; Loy, C. C.; and Liu, Z. 2022. Conditional prompt learning for vision-language models. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 16816--16825

    2022

  45. [53]

    Zhou, M.; Zhou, L.; Wang, S.; Cheng, Y.; Li, L.; Yu, Z.; and Liu, J. 2021. Uc2: Universal cross-lingual cross-modal vision-and-language pre-training. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 4155--4165

    2021

  46. [54]

    , " * 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...

  47. [55]

    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.