Pith. sign in

REVIEW 2 major objections 4 minor 51 references

DLF: Disentangled-Language-Focused Multimodal Sentiment Analysis

T0 review · 2 major / 4 minor · reviewed 2026-08-11 · deepseek-v4-flash

Pith's one-line read DLF improves multimodal sentiment analysis by disentangling shared and specific features, focusing attention on language, and predicting hierarchically, achieving top reported accuracy on CMU-MOSI and CMU-MOSEI.

desk verdict A coherent, well-ablationed MSA architecture with a language-focused cross-attention twist; the main empirical claim is only as strong as the heterogeneous baseline table, so the paper is conditionally useful and worth refereeing. read the letter →

arxiv 2412.12225 v3 pith:565HO46D submitted 2024-12-16 cs.LG cs.AIcs.CLcs.MM

classification cs.LGcs.AIcs.CLcs.MM
keywords multimodalsentimentanalysisfeaturedisentanglementlanguage-focusedattractorcross-modalattentionhierarchicalpredictiongeometricregularizationCMU-MOSICMU-MOSEI
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 the usual symmetric treatment of modalities in multimodal sentiment analysis creates redundant and conflicting information, while language is empirically the dominant modality. To exploit this, it proposes DLF, a framework that first splits each modality into shared and specific feature spaces using four geometric regularization losses, then enriches language in the specific space with a Language-Focused Attractor built on language-query cross-attention, and finally predicts from shared, specific, and fused features hierarchically. On CMU-MOSI and CMU-MOSEI, DLF reports the best accuracy on nearly every metric among eleven baselines, with the largest gains on fine-grained 7-class accuracy. If correct, the value is a concrete recipe: make the strongest modality the hub of multimodal transfer instead of treating all modalities as equal partners.

What carries the argument

The Language-Focused Attractor (LFA) is the load-bearing component: a multimodal Transformer block in which language-specific features always form the query, and all modalities supply keys and values, giving a targeted $X \to \text{Language}$ enhancement path. The disentanglement module feeds it by projecting each modality into shared and specific spaces, refined by four geometric losses: reconstruction, specific reconstruction, modified triplet, and soft orthogonality. Hierarchical predictions—shared loss, specific loss, and final fusion loss—complete the pipeline and are ablated individually.

What would settle it

Run DLF and all eleven baselines in one codebase with the same BERT/Facet/COVAREP features, the same train-validation-test splits, and several random seeds; if the MOSI Acc-7 margin over DMD (47.08 vs. 46.06) and the MOSEI Acc-2 margins shrink to within seed variance, the empirical superiority claim is falsified.

Watch

Extended reading notes

Core claim

DLF's central claim is that disentangling modality-shared from modality-specific representations, followed by a language-focused attractor and hierarchical predictions, yields better multimodal sentiment predictions than previous fusion, distillation, and disentanglement methods. In the specific space, all cross-attention branches use language as the query and vision, audio, and language as keys and values, so complementary information flows only into language rather than between all pairs. The shared space is trained with a modified triplet loss that pulls same-sentiment cross-modal samples together, and a soft orthogonality loss separates shared from specific features. The final representation concatenates enhanced specific features with a transformer-processed shared feature, and the model is trained with losses at the shared, specific, and final levels simultaneously.

Load-bearing premise

The central claim assumes that the baseline numbers copied from different papers and repositories are directly comparable to DLF's runs, since the reported one-to-four point gaps are not accompanied by error bars or significance tests.

Editorial extensions

If this is right

  • Symmetric cross-modal attention may be wasting capacity; future multimodal sentiment models can route transfer into a chosen dominant modality.
  • The four geometric disentanglement losses offer a simple, adversarial-free regularization recipe that could transfer to other multimodal tasks.
  • Supervising intermediate shared and specific representations, not just the final output, appears to improve accuracy and could stabilize training.
  • The reported confusion-matrix analysis points to the long-tailed highly-negative and highly-positive classes as the remaining bottleneck.

Reading between the lines

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

  • One testable extension not in the paper: replacing the language query with a vision or audio query would isolate whether the gains come from language dominance or merely from asymmetric attention.
  • Because each individual loss ablation changes MOSI Acc-7 by less than about 1.7 points, some of the disentanglement benefit may amount to regularization; a reader could compare against an equally strong generic regularizer.
  • The framework assumes complete modalities; applying it to missing-modality settings would require a way to supply absent keys or values, which the paper names as its own limitation.
  • The long-tailed class distribution suggests that combining DLF with class-balanced or reweighted objectives could push accuracy further beyond the reported numbers.
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

2 major / 4 minor

Summary. The paper proposes DLF, a multimodal sentiment analysis framework that disentangles shared and modality-specific features using four geometric regularization terms, enhances language representations through a Language-Focused Attractor (LFA) based on language-guided cross-attention, and combines shared, specific, and final predictions in a hierarchical loss. Experiments on CMU-MOSI and CMU-MOSEI report superior performance over eleven baselines on most metrics, with ablations on MOSI showing that each component contributes, especially LFA. The code is publicly available.

Significance. If the reported results hold, DLF offers a plausible and well-motivated way to address redundancy and conflict in multimodal fusion by explicitly routing complementary information toward the dominant language modality. The paper ships public code, and the ablation study is clean and shows that each proposed component contributes, with LFA removal causing the largest drop. The main contribution is empirical: the claimed gains over eleven baselines on two standard benchmarks. However, the strength of that claim is currently limited by the heterogeneity of the baseline sources and the absence of variance or significance information, as detailed in the major comments.

major comments (2)
  1. [Main Results, Table 1] The central empirical claim that DLF outperforms eleven baselines is not yet supported by directly comparable numbers. Baseline rows are taken from at least three different sources (THUIAR MMSA page, the MISA paper, and a reproduced DMD run), and no standard deviations, confidence intervals, or significance tests are reported for any row. The decisive margins are thin: on MOSI, DLF's MAE (0.731) is worse than MAG-BERT's (0.727), and on MOSEI, DLF's Acc-7 advantage over reproduced DMD is only 1.12 points (53.90 vs 52.78). In MSA, 1-point Acc-7 differences can easily arise from preprocessing, feature-extractor version, evaluation script, or seed differences. Because the code is public, the authors should provide a unified re-run of DLF and the closest baselines under identical conditions, or at least seed-level variance for DLF and for the baselines obtained in-house, so that the reader can assess whether the one-to-four point gaps are method effects rather than pipeline artifacts.
  2. [Eq. (7)] The modified triplet loss as written appears to have the wrong sign if d(·, ·) is the cosine similarity as stated in the text. The equation is max(0, d(S, P) − d(S, N) + µ), which, when minimized, encourages the positive pair to be less similar and the negative pair to be more similar. The standard triplet formulation with a similarity measure would be max(0, d(S, N) − d(S, P) + µ), or the authors should clarify that d denotes a distance (e.g., 1 − cosine similarity) rather than cosine similarity. Since the ablation shows that removing Lm degrades performance, this sign issue is not merely cosmetic; it must be corrected or explicitly justified in the text, and the reader should be told which form is actually implemented in the released code.
minor comments (4)
  1. [Eq. (14)] The hierarchical prediction loss L_MSA introduces L_Sh and L_Sp^m without defining them. Please specify that they are computed analogously to Eq. (13) using the shared and modality-specific predictors, or provide their explicit forms.
  2. [Eq. (10)] The notation for the cross-attention operation is confusing: the first line uses Q_L, K_m, V_m, but the second line expands Q_L as (Sp^L)W_Q^L and K_m as (Sp^m)W_K^T, with the transpose placement unclear. Please align the notation between the abstract MCA definition and the concrete instantiation.
  3. [Table 1] The footnote markers are applied inconsistently. For example, the TFN row is marked with * but the Acc-5 entry carries a †; the LMF row has a similar pattern. Clarify which source each number comes from, and consider adding a column or per-cell source indicator.
  4. [Related Work] The sentence 'Recent approaches also leverage large multimodal language models for flexible interactions (Wu et al. 2023)' cites Next-GPT, which is about any-to-any multimodal generation rather than multimodal sentiment analysis. Please adjust the citation or the claim so that the reference supports the stated context.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity identified: DLF's components and losses are independently defined and evaluated against external baselines; no prediction reduces to a fit or self-citation.

full rationale

Walking the paper's derivation chain, the method is defined by explicit equations (1)-(15): shared and specific encoders, a reconstruction loss, a specific-consistency loss, a triplet loss, an orthogonality loss, the Language-Focused Attractor with language-query cross-attention, and hierarchical prediction losses. None of these quantities is defined in terms of the reported test metric, and no hyperparameter is stated as being fitted to the test labels. The central empirical claim is benchmark superiority over eleven baselines; the baseline numbers are drawn from multiple external sources, which raises comparability and reproducibility concerns, but that is not circularity under the stated rules. The 'language is dominant' motivation is justified by external prior work and by the paper's own unimodal ablation, and even if that ablation influenced the design, it is an empirical observation rather than a definitional equivalence. The only author-associated artifact referenced is the public code repository, and no load-bearing argument reduces to a self-citation. The ablation studies removing LFA, FDM, and HP support each component's contribution empirically without equating the component to the outcome by construction. Therefore, the paper is self-contained with respect to circularity, and the appropriate score is 0.

Assumptions & free parameters 3 free parameters · 3 assumptions · 0 invented entities

The framework relies on standard attention and a set of four regularization losses. Its free parameters are loss weights and a margin, none of which are reported numerically in the paper. No new physical or conceptual entities are introduced.

free parameters (3)
  • decoupling loss weights lambda_r, lambda_s, lambda_m, lambda_o = not reported
    Used in Eq. 9 to balance reconstruction, specific, triplet, and orthogonality losses; chosen by hand or tuned on validation, not justified in paper.
  • MSA loss weights beta_f, beta_Sh, beta_Sp = not reported
    Used in Eq. 14 to weight hierarchical prediction losses; values not stated.
  • triplet margin mu = not reported
    Appears in Eq. 7; no value or selection procedure given.
assumptions (3)
  • domain assumption Language is the dominant modality for sentiment in MOSI and MOSEI.
    Motivates the LFA design; supported by Table 2 unimodal results on MOSI, but assumed transferable beyond this dataset.
  • domain assumption CMU-MOSI and CMU-MOSEI with BERT, Facet, and COVAREP features are representative and correctly preprocessed.
    All claims are evaluated on these two datasets and feature sets; no evidence they generalize.
  • standard math Standard attention and transformer building blocks work as expected in this setting.
    The MCA formulation in Eq. 10 relies on standard softmax attention.

how reviews work

0 comments
Cite this review

Pith. "Pith review of DLF: Disentangled-Language-Focused Multimodal Sentiment Analysis." pith.science (2026). https://pith.science/paper/565HO46D

@misc{pith2026241212225,
  author       = {Pith},
  title        = {Pith review of: DLF: Disentangled-Language-Focused Multimodal Sentiment Analysis},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/565HO46D}},
  note         = {Machine review of arXiv:2412.12225}
}
read the original abstract

Multimodal Sentiment Analysis (MSA) leverages heterogeneous modalities, such as language, vision, and audio, to enhance the understanding of human sentiment. While existing models often focus on extracting shared information across modalities or directly fusing heterogeneous modalities, such approaches can introduce redundancy and conflicts due to equal treatment of all modalities and the mutual transfer of information between modality pairs. To address these issues, we propose a Disentangled-Language-Focused (DLF) multimodal representation learning framework, which incorporates a feature disentanglement module to separate modality-shared and modality-specific information. To further reduce redundancy and enhance language-targeted features, four geometric measures are introduced to refine the disentanglement process. A Language-Focused Attractor (LFA) is further developed to strengthen language representation by leveraging complementary modality-specific information through a language-guided cross-attention mechanism. The framework also employs hierarchical predictions to improve overall accuracy. Extensive experiments on two popular MSA datasets, CMU-MOSI and CMU-MOSEI, demonstrate the significant performance gains achieved by the proposed DLF framework. Comprehensive ablation studies further validate the effectiveness of the feature disentanglement module, language-focused attractor, and hierarchical predictions. Our code is available at https://github.com/pwang322/DLF.

Figures

Figures reproduced from arXiv: 2412.12225 by the authors.

Figure 1
Figure 1. Task pipeline of the Multimodal Sentiment Analy [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Overview of the proposed DLF framework. The framework follows a pipeline of feature extraction, disentanglement, [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. The details of the proposed LFA. The language [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Left: Confusion matrix on MOSI. Right: Corre￾sponding accuracy for each sentiment. HN: Highly Nega￾tive; N: Negative; WN: Weakly Negative; NT: Neutral; WP: Weak Positive; P: Positive; HP: Highly Positive [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]
Figure 5
Figure 5. Figure 5: Visualization of the fused multimodal represen [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

51 extracted references · 35 canonical work pages

  1. [1]

    Ali, K.; and Hughes, C. E. 2023. A Unified Transformer-based Network for Multimodal Emotion Recognition. arXiv preprint arXiv:2308.14160

  2. [2]

    H.; Zeeshan, M

    Aslam, M. H.; Zeeshan, M. O.; Pedersoli, M.; Koerich, A. L.; Bacon, S.; and Granger, E. 2023. Privileged Knowledge Distillation for Dimensional Emotion Recognition in the Wild. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 3337--3346

  3. [3]

    Awal, R.; Ahmadi, S.; Zhang, L.; and Agrawal, A. 2024. VisMin: Visual Minimal-Change Understanding. arXiv preprint arXiv:2407.16772

  4. [4]

    Baltru s aitis, T.; Robinson, P.; and Morency, L.-P. 2016. Openface: an open source facial behavior analysis toolkit. In 2016 IEEE winter conference on applications of computer vision (WACV), 1--10. IEEE

  5. [5]

    Degottex, G.; Kane, J.; Drugman, T.; Raitio, T.; and Scherer, S. 2014. COVAREP—A collaborative voice analysis repository for speech technologies. In 2014 ieee international conference on acoustics, speech and signal processing (icassp), 960--964. IEEE

  6. [6]

    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

  7. [7]

    Ezzameli, K.; and Mahersia, H. 2023. Emotion recognition from unimodal to multimodal analysis: A review. Information Fusion, 101847

  8. [8]

    Geetha, A.; Mala, T.; Priyanka, D.; and Uma, E. 2024. Multimodal Emotion Recognition with deep learning: advancements, challenges, and future directions. Information Fusion, 105: 102218

Show all 51 references
  1. [9]

    Guan, T.; Liu, F.; Wu, X.; Xian, R.; Li, Z.; Liu, X.; Wang, X.; Chen, L.; Huang, F.; Yacoob, Y.; et al. 2024. HallusionBench: an advanced diagnostic suite for entangled language hallucination and visual illusion in large vision-language models. In Proceedings of the IEEE/CVF C...

  2. [10]

    Guo, J.; Tang, J.; Dai, W.; Ding, Y.; and Kong, W. 2022. Dynamically adjust word representations using unaligned multimodal information. In Proceedings of the 30th ACM international conference on multimedia, 3394--3402

  3. [11]

    Guo, Q.; Wang, X.; Wu, Y.; Yu, Z.; Liang, D.; Hu, X.; and Luo, P. 2020. Online knowledge distillation via collaborative learning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 11020--11029

  4. [12]

    Gupta, S.; Hoffman, J.; and Malik, J. 2016. Cross modal distillation for supervision transfer. In Proceedings of the IEEE conference on computer vision and pattern recognition, 2827--2836

  5. [13]

    Hazarika, D.; Zimmermann, R.; and Poria, S. 2020. Misa: Modality-invariant and-specific representations for multimodal sentiment analysis. In Proceedings of the 28th ACM international conference on multimedia, 1122--1131

  6. [14]

    Huang, J.; Tao, J.; Liu, B.; Lian, Z.; and Niu, M. 2020. Multimodal transformer fusion for continuous emotion recognition. In ICASSP 2020-2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 3507--3511. IEEE

  7. [15]

    Kim, D.; and Kang, P. 2022. Cross-modal distillation with audio--text fusion for fine-grained emotion classification using BERT and Wav2vec 2.0. Neurocomputing, 506: 168--183

  8. [16]

    Kim, H.; and Mnih, A. 2018. Disentangling by factorising. In International conference on machine learning, 2649--2658. PMLR

  9. [17]

    Kim, K.; and Park, S. 2023. AOBERT: All-modalities-in-One BERT for multimodal sentiment analysis. Information Fusion, 92: 37--45

  10. [18]

    Lei, Y.; Yang, D.; Li, M.; Wang, S.; Chen, J.; and Zhang, L. 2023. Text-oriented modality reinforcement network for multimodal sentiment analysis from unaligned multimodal sequences. In CAAI International Conference on Artificial Intelligence, 189--200. Springer

  11. [19]

    Li, M.; Yang, D.; Lei, Y.; Wang, S.; Wang, S.; Su, L.; Yang, K.; Wang, Y.; Sun, M.; and Zhang, L. 2024 a . A Unified Self-Distillation Framework for Multimodal Sentiment Analysis with Uncertain Missing Modalities. In Proceedings of the AAAI Conference on Artificial Intelligenc...

  12. [20]

    Li, M.; Yang, D.; Zhao, X.; Wang, S.; Wang, Y.; Yang, K.; Sun, M.; Kou, D.; Qian, Z.; and Zhang, L. 2024 b . Correlation-Decoupled Knowledge Distillation for Multimodal Sentiment Analysis with Incomplete Modalities. In Proceedings of the IEEE/CVF Conference on Computer Vision ...

  13. [21]

    Li, Y.; Wang, Y.; and Cui, Z. 2023. Decoupled Multimodal Distilling for Emotion Recognition. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 6631--6640

  14. [22]

    Li, Y.; Zeng, J.; Shan, S.; and Chen, X. 2019. Self-supervised representation learning from videos for facial action unit detection. In Proceedings of the IEEE/CVF Conference on Computer vision and pattern recognition, 10924--10933

  15. [23]

    Liang, T.; Lin, G.; Feng, L.; Zhang, Y.; and Lv, F. 2021. Attention is not enough: Mitigating the distribution discrepancy in asynchronous multimodal sequence fusion. In Proceedings of the IEEE/CVF International Conference on Computer Vision, 8148--8156

  16. [24]

    B.; Liang, P

    Liu, Z.; Shen, Y.; Lakshminarasimhan, V. B.; Liang, P. P.; Zadeh, A.; and Morency, L.-P. 2018. Efficient low-rank multimodal fusion with modality-specific factors. arXiv preprint arXiv:1806.00064

  17. [25]

    Lv, F.; Chen, X.; Huang, Y.; Duan, L.; and Lin, G. 2021. Progressive modality reinforcement for human multimodal emotion recognition from unaligned multimodal sequences. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2554--2562

  18. [26]

    Mao, H.; Yuan, Z.; Xu, H.; Yu, W.; Liu, Y.; and Gao, K. 2022. M-sena: An integrated platform for multimodal sentiment analysis. arXiv preprint arXiv:2203.12441

  19. [27]

    P.; Manzini, T.; Morency, L.-P.; and P \'o czos, B

    Pham, H.; Liang, P. P.; Manzini, T.; Morency, L.-P.; and P \'o czos, B. 2019. Found in translation: Learning robust joint representations by cyclic translations between modalities. In Proceedings of the AAAI Conference on Artificial Intelligence, 6892--6899

  20. [28]

    K.; Lee, S.; Zadeh, A.; Mao, C.; Morency, L.-P.; and Hoque, E

    Rahman, W.; Hasan, M. K.; Lee, S.; Zadeh, A.; Mao, C.; Morency, L.-P.; and Hoque, E. 2020. Integrating multimodal information in large pretrained transformers. In Proceedings of the conference. Association for Computational Linguistics. Meeting, volume 2020, 2359. NIH Public Access

  21. [29]

    Schroff, F.; Kalenichenko, D.; and Philbin, J. 2015. Facenet: A unified embedding for face recognition and clustering. In Proceedings of the IEEE conference on computer vision and pattern recognition, 815--823

  22. [30]

    Sun, L.; Lian, Z.; Liu, B.; and Tao, J. 2023. Efficient multimodal transformer with dual-level feature restoration for robust multimodal sentiment analysis. IEEE Transactions on Affective Computing, 15(1): 309--325

  23. [31]

    Thuiar . 2024. MMSA: A Multi-Modal Sentiment Analysis Toolkit. https://github.com/thuiar/MMSA. Accessed: 2024-08-10

  24. [32]

    H.; Bai, S.; Liang, P

    Tsai, Y.-H. H.; Bai, S.; Liang, P. P.; Kolter, J. Z.; Morency, L.-P.; and Salakhutdinov, R. 2019. Multimodal transformer for unaligned multimodal language sequences. In Proceedings of the conference. Association for computational linguistics. Meeting, volume 2019, 6558. NIH Pu...

  25. [33]

    H.; Liang, P

    Tsai, Y.-H. H.; Liang, P. P.; Zadeh, A.; Morency, L.-P.; and Salakhutdinov, R. 2018. Learning factorized multimodal representations. arXiv preprint arXiv:1806.06176

  26. [34]

    Wang, W.; Ding, L.; Shen, L.; Luo, Y.; Hu, H.; and Tao, D. 2024. Wisdom: Improving multimodal sentiment analysis by fusing contextual world knowledge. In Proceedings of the 32nd ACM International Conference on Multimedia, 2282--2291

  27. [35]

    Williams, J.; Comanescu, R.; Radu, O.; and Tian, L. 2018 a . Dnn multimodal fusion techniques for predicting video sentiment. In Proceedings of grand challenge and workshop on human multimodal language (Challenge-HML), 64--72

  28. [36]

    Williams, J.; Kleinegesse, S.; Comanescu, R.; and Radu, O. 2018 b . Recognizing emotions in video using multimodal dnn feature fusion. In Proceedings of Grand Challenge and Workshop on Human Multimodal Language (Challenge-HML), 11--19

  29. [37]

    Wu, S.; Fei, H.; Qu, L.; Ji, W.; and Chua, T.-S. 2023. Next-gpt: Any-to-any multimodal llm. arXiv preprint arXiv:2309.05519

  30. [38]

    Xu, P.; Zhu, X.; and Clifton, D. A. 2023. Multimodal learning with transformers: A survey. IEEE Transactions on Pattern Analysis and Machine Intelligence, 45(10): 12113--12132

  31. [39]

    Yang, D.; Chen, Z.; Wang, Y.; Wang, S.; Li, M.; Liu, S.; Zhao, X.; Huang, S.; Dong, Z.; Zhai, P.; et al. 2023 a . Context de-confounded emotion recognition. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 19005--19015

  32. [40]

    Yang, D.; Huang, S.; Kuang, H.; Du, Y.; and Zhang, L. 2022 a . Disentangled representation learning for multimodal emotion recognition. In Proceedings of the 30th ACM International Conference on Multimedia, 1642--1651

  33. [41]

    Yang, D.; Huang, S.; Liu, Y.; and Zhang, L. 2022 b . Contextual and cross-modal interaction for multi-modal speech emotion recognition. IEEE Signal Processing Letters, 29: 2093--2097

  34. [42]

    Yang, D.; Kuang, H.; Huang, S.; and Zhang, L. 2022 c . Learning modality-specific and-agnostic representations for asynchronous multimodal language sequences. In Proceedings of the 30th ACM International Conference on Multimedia, 1708--1717

  35. [43]

    Yang, K.; Yang, D.; Zhang, J.; Wang, H.; Sun, P.; and Song, L. 2023 b . What2comm: Towards communication-efficient collaborative perception via feature decoupling. In Proceedings of the 31st ACM International Conference on Multimedia, 7686--7695

  36. [44]

    Yu, W.; Xu, H.; Yuan, Z.; and Wu, J. 2021. Learning modality-specific representations with self-supervised multi-task learning for multimodal sentiment analysis. In Proceedings of the AAAI conference on artificial intelligence, 10790--10797

  37. [45]

    Zadeh, A.; Chen, M.; Poria, S.; Cambria, E.; and Morency, L.-P. 2017. Tensor fusion network for multimodal sentiment analysis. arXiv preprint arXiv:1707.07250

  38. [46]

    P.; Mazumder, N.; Poria, S.; Cambria, E.; and Morency, L.-P

    Zadeh, A.; Liang, P. P.; Mazumder, N.; Poria, S.; Cambria, E.; and Morency, L.-P. 2018 a . Memory fusion network for multi-view sequential learning. In Proceedings of the AAAI conference on artificial intelligence, 5634--5641

  39. [47]

    Zadeh, A.; Zellers, R.; Pincus, E.; and Morency, L.-P. 2016. Multimodal sentiment intensity analysis in videos: Facial gestures and verbal messages. IEEE Intelligent Systems, 31(6): 82--88

  40. [48]

    B.; Liang, P

    Zadeh, A. B.; Liang, P. P.; Poria, S.; Cambria, E.; and Morency, L.-P. 2018 b . Multimodal language analysis in the wild: Cmu-mosei dataset and interpretable dynamic fusion graph. In Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics, 2236--2246

  41. [49]

    Zhang, H.; Wang, Y.; Yin, G.; Liu, K.; Liu, Y.; and Yu, T. 2023. Learning Language-guided Adaptive Hyper-modality Representation for Multimodal Sentiment Analysis. arXiv preprint arXiv:2310.05804

  42. [50]

    , " * 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...

  43. [51]

    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.