Pith. sign in

REVIEW 3 major objections 6 minor 64 references

Teach Me How to Denoise: A Universal Framework for Denoising Multi-modal Recommender Systems via Guided Calibration

T0 review · 3 major / 6 minor · reviewed 2026-08-16 · deepseek-v4-flash

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

Pith's one-line read GUIDER is a plug-and-play denoising layer for multi-modal recommender systems: it relabels noisy feedback using text and image similarity, then distills clean ranking knowledge into the model via optimal transport.

desk verdict A useful negative result about IDRec denoising in multimodal recommenders, wrapped in a plug-and-play framework whose empirical gains are consistent but whose label construction needs direct validation. read the letter →

arxiv 2504.14214 v1 pith:2UKW5T6C submitted 2025-04-19 cs.IR

Hongji Li , Hanwen Du , Youhua Li , Junchen Fu , Chunxiao Li , Ziyi Zhuang , Jiakang Li , Yongxin Ni This is my paper · ORCID
classification cs.IR
keywords multi-modalrecommendersystemsrecommendationdenoisingknowledgedistillationoptimaltransportBayesianpersonalizedrankingmodalitysimilaritycalibrationnoisyimplicitfeedbackplug-and-playframework
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 argues that denoising methods built for ID-based recommenders break down in multi-modal recommenders, because the interaction-similarity scores they rely on no longer separate clean from noisy feedback in the multi-modal representation space. It proposes GUIDER, a plug-in training layer that reclassifies each user-item interaction using text and image similarity to items already judged clean, trains a denoised ID-based teacher with a ranking loss, and distills that teacher into the multi-modal student using an optimal-transport distance. Across four public datasets and five base recommenders, adding GUIDER improves Recall and NDCG at cutoffs 5 and 20, with the reported gains marked statistically significant, while the IDRec denoising baselines often leave the base model worse off. The payoff, if the claim holds, is a reusable denoising component that any multi-modal recommender can adopt without changing its architecture.

What carries the argument

Three components carry the argument. Adaptive Modality Similarity Calibration (AMSC) is the relabeling rule: a spurious interaction becomes clean when the maximum of its text-similarity and image-similarity to a reliable item exceeds a threshold, scaled by a cross-modal confidence score from hashing. Denoising Bayesian Personalized Ranking (DBPR) is a pairwise ranking loss that treats the calibrated clean set as positives and the calibrated noisy set as negatives. The OT-based knowledge distillation objective is an entropy-regularized optimal-transport distance between teacher and student ranking logits, solved by iterative matrix scaling; it is what lets a denoised ID teacher guide the multi-modal student without applying the denoising loss inside the multi-modal space.

What would settle it

Inject known random noisy interactions at controlled rates into the training data, run AMSC, and compare its reclassified labels against the injected labels; if AMSC's precision and recall at recovering injected noise are no better than a loss-based threshold or random assignment, the calibration claim is not supported.

Watch

Extended reading notes

Core claim

The central claim is that noisy feedback in multi-modal recommenders should not be cleaned by applying IDRec denoising losses directly to the multi-modal model. GUIDER instead starts from a loss-based partition of each user's interactions, then repairs that partition: an item initially labelled spurious is reclassified as clean if the maximum of its text-similarity and image-similarity to an already-clean item exceeds a threshold, with a cross-modal hashing confidence score weighting the decision. The repaired labels feed a Denoising Bayesian Personalized Ranking objective that trains an ID-based teacher, and the teacher's ranking logits are transferred to the multi-modal student through an entropy-regularized optimal-transport distance computed by the standard iterative entropic optimal-transport solver. The paper's empirical claim is that the full combination, not knowledge distillation alone and not direct application of AMSC and DBPR to the student, yields consistent gains over T-CE, R-CE, WBPR, and BOD across VBPR, MMGCN, BM3, DRAGON, and FREEDOM on Baby, Sports, Clothing, and MicroLens.

Load-bearing premise

The load-bearing premise is that an item whose image or text looks like an item already labelled clean for a user is itself a clean interaction for that user, even though the paper also shows visually similar items can belong to different categories.

Editorial extensions

If this is right

  • Each of the five base recommenders improves over its own no-denoising baseline when GUIDER is added, and the gains are largest on MicroLens, the dataset the paper identifies as noisiest.
  • Removing any one of AMSC, DBPR, or knowledge distillation lowers performance, and replacing the OT distance with KL divergence also lowers performance, so the three-component design is load-bearing.
  • As injected noise rises from 5% to 20%, GUIDER keeps performance from decaying as fast as the T-CE, R-CE, WBPR, and BOD baselines.
  • Because GUIDER is a training-time plugin, the same layer can be attached to future multi-modal recommenders without changing their architecture.

Reading between the lines

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

  • The paper's own finding that clean and noisy interactions have overlapping similarity scores leaves open whether AMSC's value comes from accurate noise identification or from acting as a useful reweighting prior; a direct label-recovery test would separate those two explanations.
  • The OT distillation may contribute regularization beyond denoising, since it aligns the geometry of teacher and student logits; gains could therefore partly reflect better knowledge transfer even on already-clean data.
  • A natural extension is to use AMSC's relabeling as a general noise-cleaning step for multimodal representation learning beyond recommendation, such as filtering image-text pairs before contrastive pretraining.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 6 minor

Summary. The paper proposes GUIDER, a universal denoising framework for multi-modal recommender systems (MMRecs). The method first partitions user-item interactions into reliable and spurious sets using the recommendation loss (Eq. 2), then recalibrates these sets via Adaptive Modality Similarity Calibration (AMSC) using intra-modal text/vision similarity and a cross-modal hashing confidence score (Eqs. 3-6). A teacher IDRecommender is trained with a Denoising Bayesian Personalized Ranking (DBPR) loss on the calibrated clean/noisy labels (Eq. 7), and its ranking logits are distilled into the student MMRec through an optimal-transport-based knowledge distillation objective (Eqs. 9-15). Experiments on Baby, Sports, Clothing, and MicroLens with five base models (VBPR, MMGCN, BM3, DRAGON, FREEDOM) report consistent improvements over T-CE, R-CE, WBPR, and BOD, with ablations attributing gains to each component.

Significance. If the empirical claim holds, GUIDER would provide a reusable, model-agnostic denoising layer for MMRecs, addressing a gap that the paper convincingly identifies in IDRec-oriented denoising methods. The experimental coverage is broad (5 base models × 4 datasets), the paper includes a component ablation, hyperparameter sensitivity analysis, robustness under injected noise, and statistical significance markers, and the source code is released. These are substantive strengths. However, the central label-construction mechanism is not validated against oracle noise, and the paper's own Section 3.2 findings create an internal tension with the loss-based partition that AMSC relies on. The OT-KD formulation also has technical ambiguities that affect reproducibility. The core idea is worth pursuing, but the current evidence for the correctness of the denoising labels is incomplete.

major comments (3)
  1. [§4.1, Eqs. (2)-(6)] The calibration rule in Eq. (3) is one-directional: any interaction initially placed in I_rel is permanently in I_true, and I_false is defined only as the portion of I_spr not rescued by modality similarity. No rule moves a low-loss item into the noisy set. This matters because §3.2 reports that in MMRecs the similarity distributions of clean and noisy interactions 'almost overlap' (Figure 4), which directly implies that the initial loss-based partition in Eq. (2) contains false positives. Those false positives are then frozen into the teacher's positive set for DBPR (Eq. 7) and into the KD signal. The w/o AMSC ablation (Table 3) only shows that the component changes performance; it does not show that the resulting labels are correct. Since the paper already injects oracle noise in §3.2, the authors should report precision/recall (or similar) of I_true and I_false against these oracle labels, and if the labels are found to be corrupted, the calibration rule should be revised to allow demotion of unreliable items.
  2. [§4.3, Eqs. (10)-(13)] The OT-KD objective as written is not well-defined. In Eq. (10), z is defined as a vector of log-sigmoid pairwise scores {log(σ(s_ui − s_uj))}, which are negative scalars and not a probability distribution. The optimal transport formulation in Eq. (12) imposes the marginal constraints P1_d = z_s and P^T 1_d = z_t, which require z_t and z_s to be probability histograms; the paper does not state how these logits are normalized or how the teacher and student pair sets are aligned into a common cost matrix. Independent of this, Eq. (13) contains an apparent typo: the second Sinkhorn update also uses z_s, whereas it should use z_t to enforce the teacher marginal. These issues make the central KD loss ambiguous and unreproducible, and they should be corrected and clarified.
  3. [Algorithm 1 and Eq. (2)] The source of the loss values used to form I_rel and I_spr is not specified. In Algorithm 1, the teacher is updated before the partition is computed, and both the teacher IDRec and the student MMRec have different representation spaces; the partition could differ substantially depending on which model's L_Rec is used. Since this partition, the DBPR labels, and the entire downstream KD signal all depend on this choice, the ambiguity is load-bearing for reproducibility. The paper should state explicitly which model's loss is used in Eq. (2) at each iteration.
minor comments (6)
  1. [Table 2] The row label 'MGCN' should be 'MMGCN' to match the model name in Section 5.1.3.
  2. [Table 2, VBPR + R-CE on Clothing] The R@20 value of 0.579 is implausible (R@20 cannot exceed 1 and all other values in the table are below 0.12); this appears to be a typo, likely 0.0579, and should be corrected.
  3. [§5.4.2] The default value of S_thres used in the main experiments (Table 2) is not reported; the sensitivity analysis in Figure 7 shows the effect of different thresholds but does not identify the operating point used elsewhere.
  4. [§3.2, Figure 4] The paper does not specify which base models, training epochs, or checkpoints were used to produce the similarity-score distributions in Figure 4, which limits the reader's ability to judge how representative these curves are of the actual MMRecs used in the main experiments.
  5. [References] Reference [41] contains a formatting error: the names 'Beibei Kong, Zhijin Wang, Bo Hu, and Zang Li' appear inserted into the citation text before 'TransRec', and the entry is incomplete.
  6. [Abstract and Contributions] The abstract contains a typo ('multi-modalrecommndation') and the contributions section has an ungrammatical phrase ('with via fine-grained re-calibration'); these should be fixed in revision.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: held-out recommendation metrics are external to the AMSC/DBPR pseudo-labeling loop.

full rationale

GUIDER's claimed contribution is evaluated with Recall/NDCG on held-out interactions, not with a quantity that is built into the method. The AMSC partition (Eqs. 2-3) is self-referential in the standard self-training sense: I_true is derived from the IDRec teacher's loss and then used to train that teacher via DBPR (Eq. 7), and the student is trained by distilling the teacher's logits (Eqs. 8-11). This is a feedback loop, but it does not make the reported predictions equivalent to the inputs by construction: the test metrics are computed on interactions that were not used to define I_true or I_false. No equation in the paper reduces to another equation as a tautology; the OT-based KD is a standard Sinkhorn objective, and the paper does not fit a parameter and then 'predict' that same parameter. The asymmetry noted in Eq. 3 (items in I_rel are never demoted) is a robustness/correctness concern about label quality, not circularity, and the paper's self-citations (e.g., the MicroLens dataset [30]) are to public, externally available resources rather than to an unverified self-referential theorem. The ablations (w/o KD, w/o DBPR, w/o AMSC, w/ KL) provide independent evidence that each component changes held-out performance, which is consistent with a self-contained empirical derivation.

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

The method is an empirical framework built on top of standard deep learning and optimal transport machinery. Its main burdens are per-dataset hyperparameters S_thres and lambda, the unspecified CMH projection parameters, and several domain assumptions about modality similarity being a reliable noise indicator despite the paper's own evidence to the contrary. No new physical or structural entities are introduced.

free parameters (4)
  • S_thres = Not stated; tuned per dataset (swept 0.2 to 1.0 in Figure 7)
    Modality similarity threshold in Eq. (3) that controls whether a spurious item is reclassified as clean. No default is given; Table 2 results presumably use per-dataset tuned values.
  • lambda (entropy regularization) = Not stated; swept 0.2 to 0.8 in Figure 6
    Entropy regularization coefficient in the Sinkhorn OT solution, Eq. (11) and (14). No default is given; per-dataset values are not reported.
  • CMH projection W and bias b = Not described
    Parameters of hash(x) = sign(W^T x + b) in Eq. (6). If learned, training is not specified; if random, the confidence score is essentially random. Either way, this is an untracked free component that the framework depends on.
  • Learning rate and weight decay = Search ranges [1e-4,1e-3] and [1e-4,1e-2]
    Standard optimizer hyperparameters; final values are not reported.
assumptions (5)
  • ad hoc to paper The loss-based partition at the mean loss provides a useful prior for clean versus noisy interactions in MMRecs.
    Eq. (2) splits each user's interactions using LRec. Section 3.2 shows loss is not a reliable noise indicator in the MMRec space, yet the method depends on this split as the starting point for AMSC.
  • domain assumption Intra-modality cosine similarity (max of text and image similarity) is a valid reclassification signal for clean versus noisy user-item interactions.
    Eq. (3)-(5). No direct validation against oracle noise is provided, and Figure 1 shows visually similar items from different categories, which would falsely elevate S_modal.
  • domain assumption The CMH hash function sign(W^T x + b) preserves cross-modal semantic similarity and can serve as a confidence score.
    Eq. (6). The training or initialization of W and b is not described, and no evidence shows the hash-space cosine behaves as a confidence score in this setting.
  • domain assumption Optimal transport distance is a stable, symmetric KD objective that avoids over-smoothing and model collapse.
    Section 4.3 and Table 3. The claim rests on a single ablation versus KL divergence, not on a study of collapse or over-smoothing.
  • domain assumption LightGCN is a sufficiently strong IDRec teacher for all the MMRec students used.
    Section 5.1.4. No analysis of teacher-student gap or teacher capacity is provided.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Teach Me How to Denoise: A Universal Framework for Denoising Multi-modal Recommender Systems via Guided Calibration." pith.science (2026). https://pith.science/paper/2UKW5T6C

@misc{pith2026250414214,
  author       = {Pith},
  title        = {Pith review of: Teach Me How to Denoise: A Universal Framework for Denoising Multi-modal Recommender Systems via Guided Calibration},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/2UKW5T6C}},
  note         = {Machine review of arXiv:2504.14214}
}
read the original abstract

The surge in multimedia content has led to the development of Multi-Modal Recommender Systems (MMRecs), which use diverse modalities such as text, images, videos, and audio for more personalized recommendations. However, MMRecs struggle with noisy data caused by misalignment among modal content and the gap between modal semantics and recommendation semantics. Traditional denoising methods are inadequate due to the complexity of multi-modal data. To address this, we propose a universal guided in-sync distillation denoising framework for multi-modal recommendation (GUIDER), designed to improve MMRecs by denoising user feedback. Specifically, GUIDER uses a re-calibration strategy to identify clean and noisy interactions from modal content. It incorporates a Denoising Bayesian Personalized Ranking (DBPR) loss function to handle implicit user feedback. Finally, it applies a denoising knowledge distillation objective based on Optimal Transport distance to guide the alignment from modality representations to recommendation semantics. GUIDER can be seamlessly integrated into existing MMRecs methods as a plug-and-play solution. Experimental results on four public datasets demonstrate its effectiveness and generalizability. Our source code is available at https://github.com/Neon-Jing/Guider

Figures

Figures reproduced from arXiv: 2504.14214 by the authors.

Figure 1
Figure 1. An illustration of challenges in denoising MMRecs. [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. The challenges of denoising in the multi-modal [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. Training curves of different MMRecs equipped with different denoising methods on the Sports and Baby dataset. [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Normalized similarity score distributions of IDRec and MMRec on the Clothing and Baby dataset. [PITH_FULL_IMAGE:figures/full_fig_p004_4.png]
Figure 5
Figure 5. Figure 5: Framework overview of GUIDER. where Cross-Modal Hashing (CMH) [54] is a technique that maps multi-modal representations into a common hash space to enable accurate alignment. The locality-sensitive hash function hash(x) projects multi-modal data onto a hash space where…
Figure 7
Figure 7. Figure 7: Analysis of 𝑆thres sensitivity in AMSC. 0% 5% 10% 15% 20% Noise Ratio (%) 0.000 0.005 0.010 0.015 0.020 NDCG@5 TCE+VBPR 0% 5% 10% 15% 20% Noise Ratio (%) 0.000 0.005 0.010 0.015 0.020 RCE+VBPR 0% 5% 10% 15% 20% Noise Ratio (%) 0.000 0.005 0.010 0.015 0.020 WBPR+VBPR 0%…
Figure 8
Figure 8. Figure 8: Performances with increasing noise ratios. [PITH_FULL_IMAGE:figures/full_fig_p008_8.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

64 extracted references · 15 canonical work pages

  1. [1]

    Josh Achiam, Steven Adler, Sandhini Agarwal, Lama Ahmad, Ilge Akkaya, Floren- cia Leoni Aleman, Diogo Almeida, Janko Altenschmidt, Sam Altman, Shyamal Anadkat, et al. 2023. GPT-4 technical report. arXiv preprint arXiv:2303.08774 (2023)

    arXiv 2023

  2. [2]

    Martin Arjovsky and Léon Bottou. 2017. Towards principled methods for training generative adversarial networks. arXiv preprint arXiv:1701.04862 (2017)

    arXiv 2017

  3. [3]

    Jingyuan Chen, Hanwang Zhang, Xiangnan He, Liqiang Nie, Wei Liu, and Tat- Seng Chua. 2017. Attentive collaborative filtering: Multimedia recommendation with item-and component-level attention. In Proceedings of the 40th International ACM SIGIR conference on Research and Development in Information Retrieval . 335–344

    2017

  4. [4]

    Ting Chen, Simon Kornblith, Mohammad Norouzi, and Geoffrey Hinton. 2020. A simple framework for contrastive learning of visual representations. In Interna- tional conference on machine learning . PMLR, 1597–1607

    2020

  5. [5]

    Xu Chen, Hanxiong Chen, Hongteng Xu, Yongfeng Zhang, Yixin Cao, Zheng Qin, and Hongyuan Zha. 2019. Personalized fashion recommendation with visual explanations based on multimodal attention network: Towards visually explainable recommendation. In Proceedings of the 42nd International ACM SIGIR Conference on Research and Development in Information Retrieva...

    2019

  6. [6]

    Yu Cheng, Yunzhu Pan, Jiaqi Zhang, Yongxin Ni, Aixin Sun, and Fajie Yuan. 2024. An Image Dataset for Benchmarking Recommender Systems with Raw Pixels. In Proceedings of the 2024 SIAM International Conference on Data Mining (SDM) . SIAM, 418–426

    work page 2024

  7. [7]

    Marco Cuturi. 2013. Sinkhorn distances: Lightspeed computation of optimal transport. Advances in neural information processing systems 26 (2013)

    2013

  8. [8]

    Junchen Fu, Xuri Ge, Xin Xin, Alexandros Karatzoglou, Ioannis Arapakis, Jie Wang, and Joemon M Jose. 2024. IISAN: Efficiently adapting multimodal repre- sentation for sequential recommendation with decoupled PEFT. In Proceedings of the 47th International ACM SIGIR Conference on Research and Development in Information Retrieval. 687–697

    2024

Show all 64 references
  1. [9]

    Junchen Fu, Xuri Ge, Xin Xin, Alexandros Karatzoglou, Ioannis Arapakis, Kaiwen Zheng, Yongxin Ni, and Joemon M Jose. 2024. Efficient and Effective Adaptation of Multimodal Foundation Models in Sequential Recommendation. arXiv preprint arXiv:2411.02992 (2024)

    2024 arXiv

  2. [10]

    Junchen Fu, Fajie Yuan, Yu Song, Zheng Yuan, Mingyue Cheng, Shenghui Cheng, Jiaqi Zhang, Jie Wang, and Yunzhu Pan. 2024. Exploring adapter-based transfer learning for recommender systems: Empirical studies and practical insights. In Proceedings of the 17th ACM International Co...

    2024

  3. [11]

    Zeno Gantner, Lucas Drumond, Christoph Freudenthaler, and Lars Schmidt- Thieme. 2012. Personalized ranking for non-uniformly sampled items. In Pro- ceedings of KDD Cup 2011 . PMLR, 231–247

    2012

  4. [12]

    Yunjun Gao, Yuntao Du, Yujia Hu, Lu Chen, Xinjun Zhu, Ziquan Fang, and Baihua Zheng. 2022. Self-guided learning to denoise for robust recommendation. In Proceedings of the 45th International ACM SIGIR Conference on Research and Development in Information Retrieval . 1412–1422

    2022

  5. [13]

    Xavier Glorot and Yoshua Bengio. 2010. Understanding the difficulty of training deep feedforward neural networks. Proceedings of the thirteenth international conference on artificial intelligence and statistics (2010), 249–256

    2010

  6. [14]

    Ido Guy, Naama Zwerdling, Inbal Ronen, David Carmel, and Erel Uziel. 2010. Social media recommendation based on people and tags. InProceedings of the 33rd international ACM SIGIR conference on Research and development in information retrieval. 194–201

    2010

  7. [15]

    Kaiming He, Haoqi Fan, Yuxin Wu, Saining Xie, and Ross Girshick. 2020. Mo- mentum contrast for unsupervised visual representation learning. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition . 9729–9738

    2020

  8. [16]

    Ruining He and Julian McAuley. 2016. VBPR: visual bayesian personalized ranking from implicit feedback. In Proceedings of the AAAI conference on artificial intelligence, Vol. 30

    2016

  9. [17]

    Xiangnan He, Kuan Deng, Xiang Wang, Yan Li, Yongdong Zhang, and Meng Wang. 2020. LightGCN: Simplifying and powering graph convolution network for recommendation. In Proceedings of the 43rd International ACM SIGIR conference on research and development in Information Retrieval ...

    2020

  10. [18]

    Xiangnan He, Lizi Liao, Hanwang Zhang, Liqiang Nie, Xia Hu, and Tat-Seng Chua

  11. [19]

    Xiangnan He, Lizi Liao, Hanwang Zhang, Liqiang Nie, Xia Hu, and Tat-Seng Chua. 2017. Neural collaborative filtering. In Proceedings of the 26th international conference on world wide web . 173–182

    2017

  12. [20]

    Balázs Hidasi, Alexandros Karatzoglou, Linas Baltrunas, and Domonkos Tikk

  13. [21]

    Geoffrey Hinton, Oriol Vinyals, and Jeff Dean. 2015. Distilling the knowledge in a neural network. arXiv preprint arXiv:1503.02531 (2015)

    2015 arXiv

  14. [22]

    In International Conference on Learning Representations

    Session-based Recommendations with Recurrent Neural Networks. In International Conference on Learning Representations

  15. [23]

    Yifan Hu, Yehuda Koren, and Chris Volinsky. 2008. Collaborative filtering for implicit feedback datasets. In 2008 Eighth IEEE international conference on data mining. Ieee, 263–272

    2008

  16. [24]

    Yupeng Hou, Shanlei Mu, Wayne Xin Zhao, Yaliang Li, Bolin Ding, and Ji-Rong Wen. 2022. Towards universal sequence representation learning for recommender systems. In Proceedings of the 28th ACM SIGKDD Conference on Knowledge Dis- covery and Data Mining . 585–593

    2022

  17. [25]

    Yoon Kim and Alexander M Rush. 2016. Sequence-level knowledge distillation. arXiv preprint arXiv:1606.07947 (2016)

    2016 arXiv

  18. [26]

    Wei Ji, Xiangyan Liu, An Zhang, Yinwei Wei, Yongxin Ni, and Xiang Wang. 2023. Online distillation-enhanced multi-modal transformer for sequential recommen- dation. In Proceedings of the 31st ACM International Conference on Multimedia . 955–965

    2023

  19. [27]

    Youhua Li, Hanwen Du, Yongxin Ni, Yuanqi He, Junchen Fu, Xiangyan Liu, and Qi Guo. 2024. An Empirical Study of Training ID-Agnostic Multi-modal Sequential Recommenders. arXiv preprint arXiv:2403.17372 (2024)

    2024 arXiv

  20. [28]

    Diederik P Kingma and Jimmy Ba. 2014. Adam: A method for stochastic opti- mization. arXiv preprint arXiv:1412.6980 (2014)

    2014 arXiv

  21. [29]

    Julian McAuley, Christopher Targett, Qinfeng Shi, and Anton Van Den Hengel

  22. [30]

    Youhua Li, Hanwen Du, Yongxin Ni, Pengpeng Zhao, Qi Guo, Fajie Yuan, and Xi- aofang Zhou. 2024. Multi-modality is all you need for transferable recommender systems. In 2024 IEEE 40th International Conference on Data Engineering (ICDE) . IEEE, 5008–5021

    2024

  23. [31]

    Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, et al. 2021. Learning transferable visual models from natural language supervision. In International conference on machine learni...

    2021

  24. [32]

    Aditya Ramesh, Prafulla Dhariwal, Alex Nichol, Casey Chu, and Mark Chen

  25. [33]

    Yongxin Ni, Yu Cheng, Xiangyan Liu, Junchen Fu, Youhua Li, Xiangnan He, Yongfeng Zhang, and Fajie Yuan. 2023. A content-driven micro-video recommen- dation dataset at scale. arXiv preprint arXiv:2309.15379 (2023)

    2023 arXiv

  26. [34]

    Steffen Rendle, Christoph Freudenthaler, Zeno Gantner, and Lars Schmidt-Thieme

  27. [35]

    Steffen Rendle, Walid Krichene, Li Zhang, and John Anderson. 2020. Neural collab- orative filtering vs. matrix factorization revisited. In Fourteenth ACM Conference on Recommender Systems (RecSys) . 240–248

    2020

  28. [36]

    J Ben Schafer, Joseph A Konstan, and John Riedl. 2001. E-commerce recommen- dation applications. Data mining and knowledge discovery 5 (2001), 115–153

    2001

  29. [37]

    Steffen Rendle. 2010. Factorization machines. In2010 IEEE International conference on data mining. IEEE, 995–1000

    2010

  30. [38]

    Fei Sun, Junjie Liu, Jian Wu, Chao Pei, Xiao Lin, Wenwu Sun, and Houqiang Wang. 2019. BERT4Rec: Sequential Recommendation with Bidirectional Encoder Representations from Transformer. Proceedings of the 28th ACM International Conference on Information and Knowledge Management (...

    2019

  31. [39]

    Changxin Tian, Yuexiang Xie, Yaliang Li, Nan Yang, and Wayne Xin Zhao. 2022. Learning to denoise unreliable interactions for graph collaborative filtering. In Proceedings of the 45th international ACM SIGIR conference on research and development in information retrieval. 122–132

    2022

  32. [40]

    Hugo Touvron, Thibaut Lavril, Gautier Izacard, Xavier Martinet, Marie-Anne Lachaux, Timothée Lacroix, Baptiste Rozière, Naman Goyal, Eric Hambro, Faisal Azhar, et al. 2023. LLaMa: Open and efficient foundation language models. arXiv preprint arXiv:2302.13971 (2023)

    2023 arXiv

  33. [41]

    Jie Wang, Fajie Yuan, Mingyue Cheng, Joemon M Jose, and Chenyun Yu. 2022. Beibei Kong, Zhijin Wang, Bo Hu, and Zang Li. 2022. TransRec: Learning Trans- ferable Recommendation from Mixture-of-Modality Feedback. arXiv preprint arXiv:2206.06190 (2022)

    2022

  34. [42]

    Jian Shen, Yanru Qu, Weinan Zhang, and Yong Yu. 2018. Wasserstein distance guided representation learning for domain adaptation. InProceedings of the AAAI conference on artificial intelligence , Vol. 32

    2018

  35. [43]

    Yu Wang, Xin Xin, Zaiqiao Meng, Joemon M Jose, Fuli Feng, and Xiangnan He. 2022. Learning robust recommenders through cross-model agreement. In Proceedings of the ACM Web Conference 2022 . 2015–2025

    2022

  36. [44]

    Zongwei Wang, Min Gao, Wentao Li, Junliang Yu, Linxin Guo, and Hongzhi Yin

  37. [45]

    Wei Wei, Chao Huang, Lianghao Xia, and Chuxu Zhang. 2023. Multi-modal self-supervised learning for recommendation. In Proceedings of the ACM Web Conference 2023. 790–800

    2023

  38. [46]

    Yinwei Wei, Xiang Wang, Liqiang Nie, Xiangnan He, and Tat-Seng Chua. 2020. Graph-refined convolutional network for multimedia recommendation with WSDM ’25, March 10–14, 2025, Hannover, Germany Hongji Li, et al. implicit feedback. In Proceedings of the 28th ACM international co...

    2020

  39. [47]

    Wenjie Wang, Fuli Feng, Xiangnan He, Liqiang Nie, and Tat-Seng Chua. 2021. Denoising implicit feedback for recommendation. In Proceedings of the 14th ACM international conference on web search and data mining . 373–381

    2021

  40. [48]

    Jiancan Wu, Xiang Wang, Fuli Feng, Xiangnan He, Liang Chen, Jianxun Lian, and Xing Xie. 2021. Self-supervised graph learning for recommendation. In Proceed- ings of the 44th international ACM SIGIR conference on research and development in information retrieval. 726–735

    2021

  41. [49]

    Shiwen Wu, Fei Sun, Wentao Zhang, Xu Xie, and Bin Cui. 2022. Graph neural networks in recommender systems: a survey. Comput. Surveys 55, 5 (2022), 1–37

    2022

  42. [50]

    Zixuan Yi, Xi Wang, Iadh Ounis, and Craig Macdonald. 2022. Multi-modal graph contrastive learning for micro-video recommendation. In Proceedings of the 45th International ACM SIGIR Conference on Research and Development in Information Retrieval. 1807–1811

    2022

  43. [51]

    Zheng Yuan, Fajie Yuan, Yu Song, Youhua Li, Junchen Fu, Fei Yang, Yunzhu Pan, and Yongxin Ni. 2023. Where to go next for recommender systems? id- vs. modality-based recommender models revisited. In Proceedings of the 46th International ACM SIGIR Conference on Research and Deve...

    2023

  44. [52]

    Jiaqi Zhang, Yu Cheng, Yongxin Ni, Yunzhu Pan, Zheng Yuan, Junchen Fu, Youhua Li, Jie Wang, and Fajie Yuan. 2024. NineRec: A benchmark dataset suite for evaluating transferable recommendation. IEEE Transactions on Pattern Analysis and Machine Intelligence (2024)

    2024

  45. [53]

    Yinwei Wei, Xiang Wang, Liqiang Nie, Xiangnan He, Richang Hong, and Tat-Seng Chua. 2019. MMGCN: Multi-modal graph convolution network for personalized recommendation of micro-video. In Proceedings of the 27th ACM international conference on multimedia. 1437–1445

    2019

  46. [54]

    Yan Zhen and Dit-Yan Yeung. 2012. Co-Regularized Hashing for Multimodal Data. Advances in Neural Information Processing Systems 25 (2012)

    2012

  47. [55]

    Hongyu Zhou, Xin Zhou, Zhiwei Zeng, Lingzi Zhang, and Zhiqi Shen. 2023. A comprehensive survey on multimodal recommender systems: Taxonomy, evalua- tion, and future directions. arXiv preprint arXiv:2302.04473 (2023)

    2023 arXiv

  48. [56]

    Hongyu Zhou, Xin Zhou, Lingzi Zhang, and Zhiqi Shen. 2023. Enhancing dyadic relations with homogeneous graphs for multimodal recommendation. arXiv preprint arXiv:2301.12097 (2023)

    2023 arXiv

  49. [57]

    Xin Zhou and Zhiqi Shen. 2023. A tale of two graphs: Freezing and denoising graph structures for multimodal recommendation. In Proceedings of the 31st ACM International Conference on Multimedia . 935–943

    2023

  50. [58]

    Xin Zhou, Hongyu Zhou, Yong Liu, Zhiwei Zeng, Chunyan Miao, Pengwei Wang, Yuan You, and Feijun Jiang. 2023. Bootstrap latent representations for multi-modal recommendation. In Proceedings of the ACM Web Conference 2023 . 845–854

    2023

  51. [59]

    Jiwei Zhang and Yifan Peng. 2017. Deep Cross-Modal Projection Learning for Image-Text Matching in Multimedia Recommendation Systems. In Proceedings of the ACM International Conference on Multimedia . ACM

    2017

  52. [2012]

    arXiv preprint arXiv:1205.2618 (2012)

    BPR: Bayesian personalized ranking from implicit feedback. arXiv preprint arXiv:1205.2618 (2012)

    2012 arXiv

  53. [2015]

    In Proceedings of the 38th international ACM SIGIR conference on research and development in information retrieval

    Image-based recommendations on styles and substitutes. In Proceedings of the 38th international ACM SIGIR conference on research and development in information retrieval. 43–52

  54. [2016]

    In Proceedings of the 39th International ACM SIGIR conference on Research and Development in Information Retrieval

    Fast matrix factorization for online recommendation with implicit feedback. In Proceedings of the 39th International ACM SIGIR conference on Research and Development in Information Retrieval . 549–558

  55. [2022]

    arXiv preprint arXiv:2204.06125 1, 2 (2022), 3

    Hierarchical text-conditional image generation with clip latents. arXiv preprint arXiv:2204.06125 1, 2 (2022), 3

    2022 arXiv

  56. [2023]

    InProceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining

    Efficient bi-level optimization for recommendation denoising. InProceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining . 2502–2511

Pith tools

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