Pith. sign in

REVIEW 3 major objections 6 minor 1 cited by

MTPareto: A MultiModal Targeted Pareto Framework for Fake News Detection

T0 review · 3 major / 6 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read Targeted Pareto gradient integration across three fusion levels lets a multimodal fake news detector outperform baselines, with 2.40% and 1.89% accuracy gains over its own hierarchical network.

desk verdict Useful idea, unconvincing numbers: the TPareto fusion-level gradient integration is a clean extension of Pareto MTL, but FVC suffers from near-duplicate leakage and FakeSV hyperparameters are tuned on the test set, so the headline gains are unverified. read the letter →

arxiv 2501.06764 v2 pith:3D7SLYZD submitted 2025-01-12 cs.LG

Kaiying Yan , Moyang Liu , Yukun Liu , Ruibo Fu , Zhengqi Wen , Jianhua Tao , Xuefei Liu , Guanjun Li This is my paper · ORCID
classification cs.LG
keywords multimodalfakenewsdetectionParetooptimizationgradientintegrationhierarchicalfusionvideomulti-tasklearningcross-modalconflict
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

Multimodal fake news detectors often optimize worse as they add modalities, because gradients from text, audio, and images point in conflicting directions and plain additive training degrades fusion. This paper proposes MTPareto, which treats each hierarchical fusion stage as a task with its own loss and integrates the per-stage gradients using a Targeted Pareto rule. The rule keeps the all-modal objective dominant by rescaling every gradient against the all-modal gradient, discarding intermediate gradients whose direction is too far from it and truncating their combined weight. On the FakeSV and FVC datasets, the authors report that this rule raises accuracy by 2.40% and 1.89% over the same architecture without it, and that the full framework beats prior baselines.

What carries the argument

The central object is the TPareto gradient integration rule for a hierarchical fusion network. For each fusion module shared by several level losses, the algorithm computes Pareto-optimal weights for the candidate gradients by minimizing the norm of their weighted sum, using Frank-Wolfe iterations with an analytic two-task line search, then rewrites the final gradient so the all-modal gradient's weight is forced to 1: non-all-modal gradients are included as ratios $\alpha_i^j / \alpha_{\mathrm{All}}^j$ only if their cosine with the all-modal gradient is above a threshold $\gamma$, and their total relative weight is truncated at a threshold $k$. These two thresholds — set to $0.25$ and $1$ in the experiments — are what make the integration 'targeted': they block intermediate fusion levels from dominating or pulling shared parameters away from the all-modal objective.

What would settle it

Retrain the hierarchical fusion network with and without TPareto on FakeSV while selecting $\gamma$ and $k$ only on a validation split, then report test accuracy; if the 2.40% gap shrinks to near zero or reverses, the central claim fails. The same check on FVC with thresholds fixed in advance would confirm whether the improvement transfers.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is that optimization conflicts inside a multimodal fusion network are not a fixed cost but a gradient-integration problem with a targeted remedy. The authors design a three-level hierarchical fusion network—text plus audio, then adding image, then adding extra information—and assign a classification loss to each fused level. During backpropagation, each fusion module has several candidate gradients, one per level loss; TPareto solves a Pareto weighting problem per module and then rescales all weights relative to the all-modal gradient, whose weight is fixed at 1. Intermediate-level gradients are kept only when their cosine angle to the all-modal gradient exceeds 0.25 and their summed relative weight stays below 1. The result reported is that each intermediate fusion level improves and the final all-modal classifier reaches 84.50% accuracy on FakeSV and 94.35% on FVC, compared with 82.10% and 92.46% without TPareto.

Load-bearing premise

The load-bearing premise is that the two TPareto thresholds—angle cosine 0.25 and non-all-modal weight 1—chosen by looking at accuracy on FakeSV, also work on new data and on FVC; if those thresholds were effectively tuned to FakeSV's test split, the 2.40% gain could reflect overfitting rather than a general property of the algorithm.

Editorial extensions

If this is right

  • If TPareto's reported gains are genuine, then the framework's accuracy improvement over its own no-TPareto network is 2.40 percentage points on FakeSV and 1.89 on FVC.
  • The full MTPareto framework outperforms all four baselines—TikTec, FANVN, SV-FEND, and MMAD—on both datasets across accuracy, f1, recall, and precision.
  • The FakeSV ablation shows TPareto improves every fusion level, with level-I accuracy rising to 82.84%, level-II to 84.23%, and level-III to 84.50%.
  • Because the all-modal gradient always carries weight 1, the intermediate fusion objectives can assist but never displace the final all-modal objective in shared parameters.
  • The exploration study indicates the two thresholds are tunable and that the chosen values give the best all-modal accuracy in the tested ranges.

Reading between the lines

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

  • An implication the paper leaves implicit is that the threshold pair ($\gamma=0.25$, $k=1$) was selected on FakeSV by inspecting FakeSV accuracy, so the honest generalization test is whether the same pair transfers to new data; the FVC result is suggestive but does not fully rule out tuning effects.
  • A testable extension would be to replace the fixed thresholds with a small validation-driven schedule or a soft penalty on gradient conflict, which could remove the need to hand-tune these two numbers.
  • The same all-modal-anchored Pareto integration could apply to other multi-stage fusion problems where intermediate fusion levels arise naturally, such as multimodal sentiment analysis, video captioning, or medical image fusion, all of which exhibit analogous cross-modal optimization conflicts.
  • A sharper test of the mechanism would compare TPareto against simple gradient clipping or fixed loss weighting at the same thresholds; if those reproduce the gain, the Pareto solution of the weights may not be the active ingredient.
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 / 6 minor

Summary. The paper proposes MTPareto, a framework for multimodal fake news detection from short videos. A hierarchical fusion network combines text, audio, image, and extra features through three fusion modules, and a TPareto optimization algorithm computes gradient weights via a quadratic program (Eqs. (4)-(5)) for the shared parameters of each fusion module, with all-modal gradients dominating and heuristic truncation for weight and angle conflicts. The reported experiments on FakeSV and FVC claim that TPareto improves accuracy over the same architecture without it by 2.40% and 1.89%, respectively, and that the full framework outperforms the compared baselines.

Significance. The methodological idea of treating hierarchical fusion levels as multi-task objectives and using Pareto gradient integration with a designated all-modal objective is interesting and plausibly useful. The paper's ablation results are encouraging: across Fusion Levels I-III in Table II, TPareto consistently improves over the same architecture without TPareto, and the Pareto weighting is defined by a gradient-space optimization rather than directly from final accuracies. These strengths make the central claim worth taking seriously. However, the evaluation protocol currently has two load-bearing weaknesses: potential leakage through near-duplicate videos in FVC and hyperparameter selection on the same test set used for the headline FakeSV number, so the empirical support for the claimed gains is not yet conclusive.

major comments (3)
  1. [Section IV-A and IV-A2] Section IV-A1 reports that FVC contains only 200 unique fake and 180 unique real videos, while Section IV-A2 says five-fold cross-validation is used without mentioning any grouping or deduplication. With random splits over the 6,415 rows, near-duplicate uploads of the same original video appear in both training and test, so the FVC results in Table I (94.35% accuracy; 1.89% TPareto gain) may partly measure memorization of video identity rather than generalization to unseen content. The authors should run grouped cross-validation by unique video ID or otherwise demonstrate that duplicates do not cross folds, and report those results.
  2. [Section IV-C, Tables III-IV] Section IV-C selects the angle cosine threshold (0.25) and non-all-modal weight threshold (1) by inspecting accuracy on FakeSV, and the same FakeSV test set is then used for the headline numbers in Table I. Tables III and IV show 0.5-1.0% differences across 16 grid cells, so picking the best cell can manufacture part of the claimed 2.40% improvement. The authors should select thresholds on a validation split or nested cross-validation and report test-set results for the chosen configuration, ideally with multiple seeds.
  3. [Tables I and II] All headline numbers in Tables I and II are single runs with no standard deviations, confidence intervals, or significance tests. Given the small reported margins in the exploration study and the variance typical in this setting, the improvements should be accompanied by multiple-seed statistics to support the claim of a significant improvement in Section V.
minor comments (6)
  1. [Section II-B.1] The phrase 'through concentration' should be 'through concatenation' to correctly describe the feature fusion operation.
  2. [Section III-B] The text contains typos: 'donated as' should be 'denoted as', and 'minimal site effect' should be 'minimal side effect'.
  3. [Equation (3)] The clipping notation introduced in Eq. (3) is cramped and unclear; please define the range explicitly, for example by writing the clip operation as a projection onto [0,1].
  4. [Table I] There is a formatting issue in the TikTec row for FVC f1: '77. 02' contains a stray space and should be '77.02'.
  5. [Figure 1 caption] The caption phrase 'Compared with the normal additive method leading to an unexpected deviation' is hard to parse; please rephrase to state clearly what deviation the TPareto integration prevents.
  6. [Overall] The paper does not mention code or data release; for reproducibility of the threshold choices and the data splits, the authors should provide code and exact split definitions.

Circularity Check

1 steps flagged · score 6.0 of 10

FakeSV headline accuracy is the selected cell of a test-set grid search; the TPareto improvement is partly a fitted value rather than an independent prediction.

  1. fitted input called prediction [Section IV-A2 (Experimental Setup), Section IV-C (Exploration Study), Tables I, III, IV]
    "The hyperparameters angle cosine threshold and relative weight threshold of the TPareto are set to 0.25 and 1 respectively. We can tell from Tab. III that an angle cosine threshold of 0.25 appears to be near-optimal, as it achieves level-by-level improvement and optimal all-modal performance. ... Tab. I ... Our 84.15 83.94 84.48 84.50"

    The headline FakeSV accuracy (84.50, Table I) is identical to the accuracy of the grid cell gamma=0.25 in Table III and k=1 in Table IV, the cells the exploration study declares near-optimal/proper. The exploration study reports accuracies on the FakeSV dataset with no stated validation split, and the experimental setup then fixes the hyperparameters to those selected values. So the reported 2.40% improvement over the non-TPareto model is the selected maximum of a test-set grid search, not an out-of-sample estimate of the algorithm's effect. The predicted number is the fitted number by construction.

full rationale

The core TPareto algorithm is not self-referential: the gradient weights in Equations (4)-(5) come from a quadratic program over gradients, and the Pareto integration machinery is standard external multi-task learning methodology (Sener & Koltun [23], Lin et al. [21]). No load-bearing self-citation chain is present; the only overlapping-author citation ([19]) is a related-work mention in the introduction, not an argument that forces the framework's conclusion. The real circularity is confined to the FakeSV headline: the hyperparameters gamma=0.25 and k=1 are chosen by inspecting accuracy on the FakeSV dataset (Section IV-C, Tables III-IV), and the final reported FakeSV accuracy in Table I is exactly the value at those selected cells. Thus the claimed 2.40% improvement over the non-TPareto hierarchical fusion model is partly manufactured by test-set tuning: the prediction is the selection criterion. The FVC result (1.89% improvement) uses the same hyperparameters transferred without FVC-specific tuning and therefore provides a partially independent check, though Section IV-A's statement that 6,415 FVC videos correspond to only 200 unique fake and 180 unique real videos raises a separate data-leakage concern that is an evaluation-validity issue, not circularity. On balance, one headline prediction reduces to a grid-search fit, while the underlying algorithm and the FVC evaluation retain independent content; this is partial circularity rather than total equivalence. Score 6 reflects that the FakeSV central claim is compromised by construction, but the framework as a whole is not merely a restatement of its inputs.

Assumptions & free parameters 2 free parameters · 5 assumptions · 0 invented entities

The central claim rests on two tuned hyperparameters and on the assumption that intermediate fusion losses provide useful gradient signal. The all-modal dominance rule is a design choice without derivation. No new physical or conceptual entities are introduced.

free parameters (2)
  • angle cosine threshold gamma = 0.25
    Tuned in Section IV-C exploration on FakeSV; controls which non-all-modal gradients are included.
  • non-all-modal relative weight threshold k = 1
    Tuned in Section IV-C exploration on FakeSV; truncates combined non-all-modal gradient weights.
assumptions (5)
  • standard math Pareto multi-task gradient aggregation (Sener-Koltun formulation) provides a valid combined update direction.
    The optimization in Equation (1) is taken from [23] without proof; if this formulation is invalid for hierarchical fusion, TPareto's gradients are not justified.
  • domain assumption Intermediate fusion features trained with classification losses provide useful supervision for final all-modal classification.
    Section III-B defines L1, L2, L3 and assumes these auxiliary losses benefit fusion; the paper's ablations support but do not prove this.
  • ad hoc to paper The all-modal gradient should dominate, with its weight fixed to 1.
    Equations (6)-(8) set the all-modal weight to 1 with no derivation; this is a design choice that the ablation only partially tests.
  • ad hoc to paper Hyperparameters tuned on FakeSV transfer to FVC.
    Tables III-IV tune gamma and k on FakeSV and the same values are used for FVC, an assumption not validated on FVC.
  • domain assumption Pretrained BERT, VGG19, and Wav2Vec features are sufficient representations for the task.
    Section II-A uses these extractors without comparing to alternatives or investigating whether fine-tuning changes results.

how reviews work

0 comments
Cite this review

Pith. "Pith review of MTPareto: A MultiModal Targeted Pareto Framework for Fake News Detection." pith.science (2026). https://pith.science/paper/3D7SLYZD

@misc{pith2026250106764,
  author       = {Pith},
  title        = {Pith review of: MTPareto: A MultiModal Targeted Pareto Framework for Fake News Detection},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/3D7SLYZD}},
  note         = {Machine review of arXiv:2501.06764}
}
read the original abstract

Multimodal fake news detection is essential for maintaining the authenticity of Internet multimedia information. Significant differences in form and content of multimodal information lead to intensified optimization conflicts, hindering effective model training as well as reducing the effectiveness of existing fusion methods for bimodal. To address this problem, we propose the MTPareto framework to optimize multimodal fusion, using a Targeted Pareto(TPareto) optimization algorithm for fusion-level-specific objective learning with a certain focus. Based on the designed hierarchical fusion network, the algorithm defines three fusion levels with corresponding losses and implements all-modal-oriented Pareto gradient integration for each. This approach accomplishes superior multimodal fusion by utilizing the information obtained from intermediate fusion to provide positive effects to the entire process. Experiment results on FakeSV and FVC datasets show that the proposed framework outperforms baselines and the TPareto optimization algorithm achieves 2.40% and 1.89% accuracy improvement respectively.

Figures

Figures reproduced from arXiv: 2501.06764 by the authors.

Figure 1
Figure 1. Illustration of the MTPareto framework. II. MULTIMODAL HIERARCHICAL FUSION ARCHITECTURE A. Overview of Model Architecture We use pre-trained BERT, VGG19, and Wav2Vec to extract text, visual, and audio feature embeddings respectively and the number of likes is used to weight comment data. As shown in [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. A New Dataset and Benchmark for Grounding Multimodal Misinformation

    cs.SI 2025-09 conditional novelty 6.0 of 10

    GroundLie360 is a 2,044-video Snopes-derived benchmark with fine-grained annotations localizing six types of multimodal misinformation; the VLM prompting baseline FakeMark shows the task remains difficult.

Reference graph

Works this paper leans on

33 extracted references · 23 canonical work pages · cited by 1 Pith paper

  1. [1]

    J. M. Jordan, The Rise of the Algorithms: How YouTube and TikTok Conquered the World. Penn State Press, 2024

    work page 2024

  2. [2]

    The (minimal) persuasive advantage of political video over text,

    C. Wittenberg, B. M. Tappin, A. J. Berinsky, and D. G. Rand, “The (minimal) persuasive advantage of political video over text,”Proceedings of the National Academy of Sciences , vol. 118, no. 47, p. e2114388118, 2021

    work page 2021

  3. [3]

    Beyond news contents: The role of social context for fake news detection,

    K. Shu, S. Wang, and H. Liu, “Beyond news contents: The role of social context for fake news detection,” in Proceedings of the twelfth ACM international conference on web search and data mining , 2019, pp. 312–320

    work page 2019

  4. [4]

    Nlp-based feature extraction for the detection of covid-19 misinformation videos on youtube,

    J. C. M. Serrano, O. Papakyriakopoulos, and S. Hegelich, “Nlp-based feature extraction for the detection of covid-19 misinformation videos on youtube,” in Proceedings of the 1st Workshop on NLP for COVID-19 at ACL 2020 , 2020

    work page 2020

  5. [5]

    Exposing deep fakes using inconsistent head poses,

    X. Yang, Y . Li, and S. Lyu, “Exposing deep fakes using inconsistent head poses,” in ICASSP 2019-2019 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) . IEEE, 2019, pp. 8261–8265

    work page 2019

  6. [6]

    Mvae: Multimodal variational autoencoder for fake news detection,

    D. Khattar, J. S. Goud, M. Gupta, and V . Varma, “Mvae: Multimodal variational autoencoder for fake news detection,” in The world wide web conference, 2019, pp. 2915–2921

    2019

  7. [7]

    Hierarchical multi-modal contextual attention network for fake news detection,

    S. Qian, J. Wang, J. Hu, Q. Fang, and C. Xu, “Hierarchical multi-modal contextual attention network for fake news detection,” in Proceedings of the 44th international ACM SIGIR conference on research and development in information retrieval , 2021, pp. 153–162

    2021

  8. [8]

    Spotfake+: A multimodal framework for fake news detection via transfer learning (student abstract),

    S. Singhal, A. Kabra, M. Sharma, R. R. Shah, T. Chakraborty, and P. Kumaraguru, “Spotfake+: A multimodal framework for fake news detection via transfer learning (student abstract),” in Proceedings of the AAAI conference on artificial intelligence , vol. 34, no. 10, 2020, pp. 13 915–13 916

    work page 2020

Show all 33 references
  1. [9]

    Detecting fake news by exploring the consistency of multimodal data,

    J. Xue, Y . Wang, Y . Tian, Y . Li, L. Shi, and L. Wei, “Detecting fake news by exploring the consistency of multimodal data,” Information Processing & Management , vol. 58, no. 5, p. 102610, 2021

    2021

  2. [10]

    Leveraging intra and inter modality relationship for multimodal fake news detection,

    S. Singhal, T. Pandey, S. Mrig, R. R. Shah, and P. Kumaraguru, “Leveraging intra and inter modality relationship for multimodal fake news detection,” in Companion Proceedings of the Web Conference 2022, 2022, pp. 726–734

    2022

  3. [11]

    Modeling both intra-and inter- modality uncertainty for multimodal fake news detection,

    L. Wei, D. Hu, W. Zhou, and S. Hu, “Modeling both intra-and inter- modality uncertainty for multimodal fake news detection,” IEEE Trans- actions on Multimedia , 2023

    2023

  4. [12]

    Entity-oriented multi- modal alignment and fusion network for fake news detection,

    P. Li, X. Sun, H. Yu, Y . Tian, F. Yao, and G. Xu, “Entity-oriented multi- modal alignment and fusion network for fake news detection,” IEEE Transactions on Multimedia , vol. 24, pp. 3455–3468, 2021

    2021

  5. [13]

    A multimodal misinfor- mation detector for covid-19 short videos on tiktok,

    L. Shang, Z. Kou, Y . Zhang, and D. Wang, “A multimodal misinfor- mation detector for covid-19 short videos on tiktok,” in 2021 IEEE international conference on big data (big data) . IEEE, 2021, pp. 899– 908

    2021

  6. [14]

    Video rumor classification based on multi-modal theme and keyframe fusion,

    J. You, Y . Lin, D. Lin, and D. Cao, “Video rumor classification based on multi-modal theme and keyframe fusion,” in CCF Conference on Com- puter Supported Cooperative Work and Social Computing . Springer, 2022, pp. 58–72

    2022

  7. [15]

    A transformer-encoder-based multimodal multi-attention fusion network for sentiment analysis,

    C. Liu, Y . Wang, and J. Yang, “A transformer-encoder-based multimodal multi-attention fusion network for sentiment analysis,” Applied Intelli- gence, vol. 54, no. 17, pp. 8415–8441, 2024

    2024

  8. [16]

    Fakesv: A multimodal benchmark with rich social context for fake news detection on short video platforms,

    P. Qi, Y . Bu, J. Cao, W. Ji, R. Shui, J. Xiao, D. Wang, and T.-S. Chua, “Fakesv: A multimodal benchmark with rich social context for fake news detection on short video platforms,” in Proceedings of the AAAI Conference on Artificial Intelligence , vol. 37, no. 12, 2023, pp. 14...

    2023

  9. [17]

    Mitigating world biases: A multimodal multi-view debiasing framework for fake news video detection,

    Z. Zeng, M. Luo, X. Kong, H. Liu, H. Guo, H. Yang, Z. Ma, and X. Zhao, “Mitigating world biases: A multimodal multi-view debiasing framework for fake news video detection,” in ACM Multimedia 2024

    2024

  10. [18]

    Bad actor, good advisor: Exploring the role of large language models in fake news detection,

    B. Hu, Q. Sheng, J. Cao, Y . Shi, Y . Li, D. Wang, and P. Qi, “Bad actor, good advisor: Exploring the role of large language models in fake news detection,” in Proceedings of the AAAI Conference on Artificial Intelligence, vol. 38, no. 20, 2024, pp. 22 105–22 113

    2024

  11. [19]

    Fake news detection and manipulation reasoning via large vision-language models,

    R. Jin, R. Fu, Z. Wen, S. Zhang, Y . Liu, and J. Tao, “Fake news detection and manipulation reasoning via large vision-language models,” arXiv preprint arXiv:2407.02042, 2024

    2024 arXiv

  12. [20]

    Mis- information detection in social media video posts,

    K. Wang, D. Chan, S. Z. Zhao, J. Canny, and A. Zakhor, “Mis- information detection in social media video posts,” arXiv preprint arXiv:2202.07706, 2022

    2022 arXiv

  13. [21]

    Pareto multi-task learning,

    X. Lin, H.-L. Zhen, Z. Li, Q.-F. Zhang, and S. Kwong, “Pareto multi-task learning,” Advances in neural information processing systems , vol. 32, 2019

    2019

  14. [22]

    Efficient continuous pareto exploration in multi-task learning,

    P. Ma, T. Du, and W. Matusik, “Efficient continuous pareto exploration in multi-task learning,” in International Conference on Machine Learn- ing. PMLR, 2020, pp. 6522–6531

    2020

  15. [23]

    Multi-task learning as multi-objective opti- mization,

    O. Sener and V . Koltun, “Multi-task learning as multi-objective opti- mization,” Advances in neural information processing systems , vol. 31, 2018

    2018

  16. [24]

    Pareto manifold learning: Tackling multiple tasks via ensembles of single-task models,

    N. Dimitriadis, P. Frossard, and F. Fleuret, “Pareto manifold learning: Tackling multiple tasks via ensembles of single-task models,” in Interna- tional Conference on Machine Learning. PMLR, 2023, pp. 8015–8052

    2023

  17. [25]

    Pembot: Pareto- ensembled multi-task boosted trees,

    G. Swamy, A. Saladi, A. Das, and S. Niranjan, “Pembot: Pareto- ensembled multi-task boosted trees,” in Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining , 2024, pp. 5752–5761

    2024

  18. [26]

    Mmpareto: Boosting multimodal learning with innocent unimodal assistance,

    Y . Wei and D. Hu, “Mmpareto: Boosting multimodal learning with innocent unimodal assistance,” arXiv preprint arXiv:2405.17730 , 2024

    2024 arXiv

  19. [27]

    Attention is all you need,

    A. Vaswani, “Attention is all you need,” arXiv preprint arXiv:1706.03762, 2017

    2017 arXiv

  20. [28]

    Vilbert: Pretraining task-agnostic visiolinguistic representations for vision-and-language tasks,

    J. Lu, D. Batra, D. Parikh, and S. Lee, “Vilbert: Pretraining task-agnostic visiolinguistic representations for vision-and-language tasks,” Advances in neural information processing systems , vol. 32, 2019

    2019

  21. [29]

    Gated multimodal units for information fusion,

    J. Arevalo, T. Solorio, M. Montes-y G ´omez, and F. A. Gonz ´alez, “Gated multimodal units for information fusion,” arXiv preprint arXiv:1702.01992, 2017

    2017 arXiv

  22. [30]

    Revisiting frank-wolfe: Projection-free sparse convex opti- mization,

    M. Jaggi, “Revisiting frank-wolfe: Projection-free sparse convex opti- mization,” in International conference on machine learning . PMLR, 2013, pp. 427–435

    2013

  23. [31]

    A corpus of debunked and verified user-generated videos,

    O. Papadopoulou, M. Zampoglou, S. Papadopoulos, and I. Kompatsiaris, “A corpus of debunked and verified user-generated videos,” Online information review, vol. 43, no. 1, pp. 72–88, 2019

    2019

  24. [32]

    Using topic modeling and adversarial neural networks for fake news video detection,

    H. Choi and Y . Ko, “Using topic modeling and adversarial neural networks for fake news video detection,” inProceedings of the 30th ACM international conference on information & knowledge management , 2021, pp. 2950–2954

    2021

  25. [33]

    Cross-modal contrastive learning for multimodal fake news detection,

    L. Wang, C. Zhang, H. Xu, Y . Xu, X. Xu, and S. Wang, “Cross-modal contrastive learning for multimodal fake news detection,” in Proceedings of the 31st ACM international conference on multimedia , 2023, pp. 5696–5704

    2023

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.