Hierarchical MoE: Continuous Multimodal Emotion Recognition with Incomplete and Asynchronous Inputs

Guanxuan Jiang; Lei Han; Pengyuan Zhou; Yitong Zhu; Yuyang Wang

REVIEW 3 major objections 3 minor 50 references

Hierarchical MoE: Continuous Multimodal Emotion Recognition with Incomplete and Asynchronous Inputs

T0 review · 3 major / 3 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read A hierarchical mixture-of-experts model keeps continuous emotion prediction accurate when modalities are missing or out of sync, and reports state-of-the-art results on DEAP and DREAMER.

desk verdict A plausible new MoE design with a strong empirical claim the abstract doesn't let you verify. read the letter →

arxiv 2508.02133 v4 pith:DA3KHFW6 submitted 2025-08-04 cs.HC

Yitong Zhu , Lei Han , Guanxuan Jiang , PengYuan Zhou , Yuyang Wang This is my paper

classification cs.HC

keywords multimodalemotionrecognitionmixture-of-expertsmissingmodalitiesasynchronousinputscontinuousregressioncross-modalalignmentDEAPDREAMER

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

Continuous multimodal emotion recognition usually assumes every sensor stream is present and synchronized, but real deployments have dropouts and delays. This paper proposes Hi-MoE, a hierarchical mixture-of-experts model that handles missing or asynchronous inputs by soft-routing around whatever modalities are available, then using differential-attention routing over emotion prototypes to produce fine-grained continuous predictions. A cross-modal alignment module explicitly corrects temporal shifts and semantic mismatches between streams. On the DEAP and DREAMER benchmarks, the authors report state-of-the-art continuous emotion regression and robustness under simulated modality absence and asynchrony. If the results hold, affect-aware systems could rely on variable, imperfect sensor sets without a separate imputation or synchronization stage.

What carries the argument

The load-bearing mechanism is the Hierarchical Mixture-of-Experts (Hi-MoE) architecture. Its first layer, the Modality Expert Bank, uses soft routing—learned gating that distributes weight across available modality-specific experts—so that a missing stream simply receives little or no weight rather than derailing fusion. Its second layer, the Emotion Expert Bank, applies differential-attention routing over emotion prototypes, letting the model concentrate on the emotional pattern most relevant to the current input. An explicit cross-modal alignment module operates on the features before fusion to correct temporal offsets and semantic inconsistencies. Together, these components are what let the model regress continuous affect without requiring complete, synchronized input.

What would settle it

Evaluate the trained model on a test set where missingness and asynchrony are generated by a different mechanism than training—for example, dropout that is correlated with emotional intensity or with a specific sensor type, or streams shifted by delays outside the training range—and check whether continuous emotion regression accuracy degrades to baseline; if it does, the reported robustness is a fitted property of the protocol.

Watch

Extended reading notes

Core claim

The paper's central claim is that a two-level expert architecture can maintain accurate continuous emotion predictions even when input modalities are incomplete or temporally misaligned. At the first level, a Modality Expert Bank with soft routing assigns weight to each available channel and effectively ignores absent ones, so fusion does not depend on all modalities being present. At the second level, an Emotion Expert Bank uses differential-attention routing to attend to emotional prototypes, which the authors argue gives finer-grained emotion representation than fixed fusion. A cross-modal alignment module handles time shifts and semantic mismatches between modalities. The reported experiments on DEAP and DREAMER show state-of-the-art performance in continuous valence and arousal regression, with robustness to dynamic modality absence and asynchronous sampling.

Load-bearing premise

The robustness claims assume that the simulated modality absence and asynchrony used in training and testing faithfully reproduces how real sensor streams actually drop and lag; if that simulation is unrealistic, or if training and evaluation use the same masking protocol, the reported robustness may not transfer to deployment.

Editorial extensions

If this is right

If the reported results hold, emotion-recognition systems can be built without assuming all recording channels are always present; the model can operate on whatever subset of sensors is online.
The explicit alignment module implies that separate offline synchronization of physiological and audiovisual streams may become unnecessary for continuous affect prediction.
The dual-bank design separates the decision of which modalities to trust from the decision of which emotion pattern is active, which could make failures easier to diagnose.
State-of-the-art results on both DEAP and DREAMER suggest the approach transfers across datasets with different subject populations and signal types.

Reading between the lines

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

A natural extension the paper does not report is to train and test under different missingness distributions; that experiment would separate genuine robustness from adaptation to a fixed masking schedule.
The emotion prototypes in the second expert bank could be inspected to see whether they correspond to interpretable affective categories, potentially making the model useful for explaining why a particular valence or arousal score was predicted.
The same hierarchical routing design could be carried over to other asynchronous multimodal tasks, such as stress monitoring from wearables or speech- and video-based depression severity estimation, where sensor dropout is routine.
If soft routing weights are interpretable, the architecture could also serve as a sensor-quality monitor, flagging when a stream is too unreliable to trust.

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.

Desk Editor's Note

read the letter

Useful architecture, unverifiable headline claim. The dual-layer MoE (modality expert bank with soft routing, emotion expert bank with differential-attention routing, plus cross-modal alignment) is a specific new combination that directly targets sensor dropout and desynchronization — a real problem in HCI. The paper is clearly written and positioned against methods that assume complete data, which is the right baseline. It picks standard benchmarks, DEAP and DREAMER, so results are comparable with the literature.

The soft spots are all about the abstract leaving the central claim uncheckable. State-of-the-art is asserted without error bars, significance tests, baseline names, or any experimental details. The robustness claim — the reason to care — depends on a simulation of modality absence and asynchrony that is not described. If the masking and delay distributions used at train time match those at test time, the 'robustness' could be a fitted response to that protocol rather than a general capacity. That is not a demonstrated flaw, but it's the first thing I would ask for in the full paper. I'd also want to see code and data; none are promised.

A minor point: 'exceptional robustness' and 'significantly advances' are typical overclaims; ignore the adjectives.

This is a paper for the affective computing / HCI community. The architecture is concrete enough to reproduce and the empirical claim is checkable if the experimental section describes the simulation protocol. That makes it worth a serious referee. I'd recommend sending to peer review with a request for full experimental details, especially the missingness and asynchrony protocol, error bars, and ablations on routing design. The reader's low confidence is appropriate given the abstract-only scope, and the stress-test concern is well-aimed; we just can't resolve it without the methods.

Referee Report

3 major / 3 minor

Summary. The manuscript proposes Hi-MoE, a hierarchical mixture-of-experts framework for continuous multimodal emotion recognition. The architecture consists of a Modality Expert Bank with soft routing for handling missing modalities, an Emotion Expert Bank with differential-attention routing over emotional prototypes, and a cross-modal alignment module for temporal and semantic inconsistencies. The authors claim state-of-the-art performance on DEAP and DREAMER, with exceptional robustness to dynamic modality absence and asynchronous sampling. The submitted text contains only the abstract; no experimental protocol, quantitative results, or implementation details are available for verification.

Significance. If the claimed results are reproducible and the robustness protocol is realistic, the Hi-MoE architecture would be a meaningful contribution to multimodal emotion recognition, a field where incomplete and asynchronous sensor streams are common. The dual-layer expert design and explicit cross-modal alignment are reasonable architectural innovations. However, the significance can only be assessed provisionally: the empirical claim is the core of the paper, and the abstract provides no numbers, no baselines, no error bars, and no description of how missingness and asynchrony were simulated. The contribution therefore cannot currently be distinguished from a method that merely overfits to a specific, undisclosed augmentation scheme.

major comments (3)

[Abstract] The sentence 'Extensive experiments on benchmark datasets DEAP and DREAMER demonstrate our model's state-of-the-art performance' is unsupported by any quantitative data. A state-of-the-art claim requires concrete performance metrics (e.g., RMSE, CCC, PCC), the specific comparison baselines, and a description of the evaluation protocol. Without these, the claim is not verifiable, especially because the abstract offers no confidence intervals, standard deviations, or significance tests.
[Abstract] The robustness claim regarding 'dynamic modality absence and asynchronous sampling' is load-bearing but completely unspecified. The abstract does not state how modality absence and asynchrony are generated (e.g., random dropout rates, delay distributions, temporal offsets), what ranges are tested, or whether the training and evaluation protocols use the same simulation. If the model is trained and tested under identical masking and delay distributions, the reported robustness may be a learned response to that specific simulation rather than a general capacity for real-world sensor dropout and desynchronization. This protocol description is essential and must be added.
[Abstract] The phrase 'exceptional robustness' suggests a comparative evaluation, but no comparison is presented. The abstract should identify the baseline methods used under missing and asynchronous conditions and report the relative improvements. Without such comparisons, the central claim is not falsifiable from the provided text.

minor comments (3)

[Abstract] The abbreviation 'Hi-MoE' is introduced but the full term is only partially given; please define 'MoE' (Mixture-of-Experts) explicitly at first use.
[Abstract] The term 'continuous emotion regression' should be clarified in terms of target variables (e.g., valence and arousal) and the evaluation metric, since this determines the interpretability of the claimed state-of-the-art performance.
[Abstract] The phrase 'differential-attention routing to flexibly attend to emotional prototypes' is not self-explanatory; a brief definition of 'emotional prototypes' and the routing mechanism would improve readability.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity identified: the available abstract presents an empirical benchmark claim with no derivation chain that could reduce to its inputs.

full rationale

The paper's available text (the abstract) contains no derivation chain, no equations, and no citations; its central claim is empirical state-of-the-art performance on external benchmarks DEAP and DREAMER. Because evaluation is against held-out benchmark data, the prediction is not constructed from its inputs. The robustness claim mentions dynamic modality absence and asynchronous sampling, but the generation protocol is not described; this is an information gap that bears on validity and generalizability, not a demonstrated circularity. No self-definitional, fitted-input-as-prediction, or self-citation-load-bearing step can be quoted, so the circularity score is 0.

Assumptions & free parameters 4 free parameters · 4 assumptions · 1 invented entities

Abstract-only review: this ledger lists the premises the abstract's claims visibly depend on and the free-parameter classes typical of this architecture, whose values are not reported in the abstract. Each entry should be re-checked against the full text when it becomes available.

free parameters (4)

Number of experts per bank
Architecture hyperparameters for the modality and emotion expert banks; chosen by hand and not reported in the abstract.
Routing gating parameters
Soft routing and differential-attention routing require temperature or scaling choices that are tuned on validation data; values not visible in the abstract.
Cross-modal alignment loss weight
The alignment module's contribution to the total loss is weighted by a hyperparameter not reported in the abstract.
Modality missingness and asynchrony simulation parameters
The robustness training and evaluation presumably mask and delay modalities under some protocol; the protocol and its rates are not described in the abstract.

assumptions (4)

domain assumption Self-reported continuous valence and arousal labels are valid regression targets
The SOTA regression claim is measured against subjective self-reports in DEAP and DREAMER; label reliability is not discussed in the abstract.
ad hoc to paper Simulated modality incompleteness and asynchrony faithfully represent real-world conditions
The abstract extrapolates from benchmark simulation to complex real-world environments without describing how missingness and asynchrony were generated.
domain assumption The remaining modalities carry enough signal to compensate for missing ones
Soft routing recovers emotion estimates from whichever modalities are present, which presumes sufficient redundancy in the signals.
standard math Standard deep learning machinery behaves as published
Attention, MoE routing, and backpropagation are taken as reliable tools; no formal verification is offered.

invented entities (1)

Emotional prototypes
purpose: Learned reference embeddings that the emotion expert bank attends to via differential-attention routing for fine-grained emotion representation
The abstract introduces these latent reference points without any external or falsifiable grounding outside the model itself.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Hierarchical MoE: Continuous Multimodal Emotion Recognition with Incomplete and Asynchronous Inputs." pith.science (2026). https://pith.science/paper/DA3KHFW6

@misc{pith2026250802133,
  author       = {Pith},
  title        = {Pith review of: Hierarchical MoE: Continuous Multimodal Emotion Recognition with Incomplete and Asynchronous Inputs},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/DA3KHFW6}},
  note         = {Machine review of arXiv:2508.02133}
}

read the original abstract

Multimodal emotion recognition (MER) is crucial for human-computer interaction, yet real-world challenges like dynamic modality incompleteness and asynchrony severely limit its robustness. Existing methods often assume consistently complete data or lack dynamic adaptability. To address these limitations, we propose a novel Hi-MoE~(Hierarchical Mixture-of-Experts) framework for robust continuous emotion prediction. This framework employs a dual-layer expert structure. A Modality Expert Bank utilizes soft routing to dynamically handle missing modalities and achieve robust information fusion. A subsequent Emotion Expert Bank leverages differential-attention routing to flexibly attend to emotional prototypes, enabling fine-grained emotion representation. Additionally, a cross-modal alignment module explicitly addresses temporal shifts and semantic inconsistencies between modalities. Extensive experiments on benchmark datasets DEAP and DREAMER demonstrate our model's state-of-the-art performance in continuous emotion regression, showcasing exceptional robustness under challenging conditions such as dynamic modality absence and asynchronous sampling. This research significantly advances the development of intelligent emotion systems adaptable to complex real-world environments.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

50 extracted references · 42 canonical work pages

[1]

Aktürk, T.; de Graaf, T.; Abra, Y.; Şahoğlu Göktaş, S.; Özkan, D.; Kula, A.; and Güntekin, B. 2021. Event-related EEG oscillatory responses elicited by dynamic facial expression. BioMedical Engineering OnLine, 20

work page 2021
[2]

Alswaidan, N.; Menai, M.; Alswaidan, N.; and Menai, M. 2020. A survey of state-of-the-art approaches for emotion recognition in text. Knowledge and Information Systems, 62: 2937--2987

work page 2020
[3]

ArunaGladys, A.; Vetriselvi, V.; ArunaGladys, A.; and Vetriselvi, V. 2023. Survey on multimodal approaches to emotion recognition. Neurocomputing, 556: 126693

work page 2023
[4]

A.V., G.; T., M.; D., P.; and E., U. 2023. Multimodal Emotion Recognition with Deep Learning: Advancements, challenges, and future directions. Inf. Fusion, 105: 102218

work page 2023
[5]

Binias, B.; Myszor, D.; Palus, H.; and Cyran, K. 2020. Prediction of Pilot's Reaction Time Based on EEG Signals. Frontiers in Neuroinformatics, 14

work page 2020
[6]

Cheng, Y.; Li, Y.; He, J.; and Feng, R. 2024. Mixtures of experts for audio-visual learning. Advances in Neural Information Processing Systems, 37: 219--243

work page 2024
[7]

Dai, D.; Dong, L.; Ma, S.; Zheng, B.; Sui, Z.; Chang, B.; and Wei, F. 2022. S table M o E : Stable Routing Strategy for Mixture of Experts. In Muresan, S.; Nakov, P.; and Villavicencio, A., eds., Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), 7085--7095. Dublin, Ireland: Association for Com...

work page 2022
[8]

Fang, Y.; Huang, W.; Wan, G.; Su, K.; and Ye, M. 2025. EMOE: Modality-Specific Enhanced Dynamic Emotion Experts. In Proceedings of the Computer Vision and Pattern Recognition Conference, 14314--14324

work page 2025

Show all 50 references

[9]

Fedus, W.; Zoph, B.; and Shazeer, N. 2022. Switch transformers: Scaling to trillion parameter models with simple and efficient sparsity. Journal of Machine Learning Research, 23(120): 1--39

2022
[10]

T.; and Xu, X

Gao, Z.; Hu, D.; Jiang, X.; Lu, H.; Shen, H. T.; and Xu, X. 2024. Enhanced Experts with Uncertainty-Aware Routing for Multimodal Sentiment Analysis. In Proceedings of the 32nd ACM International Conference on Multimedia, MM '24, 9650–9659. New York, NY, USA: Association for Com...

2024
[11]

Han, W.; Chen, H.; and Poria, S. 2021. Improving Multimodal Fusion with Hierarchical Mutual Information Maximization for Multimodal Sentiment Analysis. In Moens, M.-F.; Huang, X.; Specia, L.; and Yih, S. W.-t., eds., Proceedings of the 2021 Conference on Empirical Methods in N...

2021
[12]

Huai, T.; Zhou, J.; Wu, X.; Chen, Q.; Bai, Q.; Zhou, Z.; and He, L. 2025. CL-MoE: Enhancing Multimodal Large Language Model with Dual Momentum Mixture-of-Experts for Continual Visual Question Answering. In Proceedings of the Computer Vision and Pattern Recognition Conference, ...

2025
[13]

Huang, J.; Li, Y.; Tao, J.; Lian, Z.; Wen, Z.; Yang, M.; and Yi, J. 2017. Continuous Multimodal Emotion Prediction Based on Long Short Term Memory Recurrent Neural Network. AVEC '17, 11–18. New York, NY, USA: Association for Computing Machinery. ISBN 9781450355025

2017
[14]

A.; Jordan, M

Jacobs, R. A.; Jordan, M. I.; Nowlan, S. J.; and Hinton, G. E. 1991. Adaptive mixtures of local experts. Neural computation, 3(1): 79--87

1991
[15]

I.; and Jacobs, R

Jordan, M. I.; and Jacobs, R. A. 1994. Hierarchical mixtures of experts and the EM algorithm. Neural computation, 6(2): 181--214

1994
[16]

A.; and Gasteratos, A

Kansizoglou, I.; Misirlis, E.; Tsintotas, K. A.; and Gasteratos, A. 2022. Continuous Emotion Recognition for Long-Term Behavior Modeling through Recurrent Neural Networks. Technologies

2022
[17]

Katsigiannis, S.; and Ramzan, N. 2018. DREAMER: A Database for Emotion Recognition Through EEG and ECG Signals From Wireless Low-cost Off-the-Shelf Devices. IEEE Journal of Biomedical and Health Informatics, 22(1): 98--107

2018
[18]

Koelstra, S.; Muhl, C.; Soleymani, M.; Lee, J.-S.; Yazdani, A.; Ebrahimi, T.; Pun, T.; Nijholt, A.; and Patras, I. 2012. DEAP: A Database for Emotion Analysis ;Using Physiological Signals. IEEE Transactions on Affective Computing, 3(1): 18--31

2012
[19]

Lawhern, V.; Solon, A.; Waytowich, N.; Gordon, S.; Hung, C.; and Lance, B. 2018. EEGNet: a compact convolutional neural network for EEG-based brain–computer interfaces. Journal of Neural Engineering, 15

2018
[20]

Lepikhin, D.; Lee, H.; Xu, Y.; Chen, D.; Firat, O.; Huang, Y.; Krikun, M.; Shazeer, N.; and Chen, Z. 2020. Gshard: Scaling giant models with conditional computation and automatic sharding. arXiv preprint arXiv:2006.16668

2020 arXiv
[21]

Li, J.; Chen, N.; Zhu, H.; Li, G.; Xu, Z.; and Chen, D. 2024. Incongruity-aware multimodal physiology signals fusion for emotion recognition. Information Fusion, 105: 102220

2024
[22]

Li, Q.; Liu, Y.; Yan, F.; Zhang, Q.; and Liu, C. 2023. Emotion recognition based on multiple physiological signals. Biomedical Signal Processing and Control, 85: 104989

2023
[23]

Li, X. 2023. Tacoformer: Token-channel compounded cross attention for multimodal emotion recognition. arXiv preprint arXiv:2306.13592

2023 arXiv
[24]

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

2021
[25]

Mittal, T.; Bhattacharya, U.; Chandra, R.; Bera, A.; and Manocha, D. 2020. M3er: Multiplicative multimodal emotion recognition using facial, textual, and speech cues. In Proceedings of the AAAI conference on artificial intelligence, volume 34, 1359--1367

2020
[26]

Muukkonen, I.; and Salmela, V. R. 2022. Representational structure of fMRI/EEG responses to dynamic facial expressions. NeuroImage, 263: 119631

2022
[27]

E.; Abdar, M.; Yen, N

Nemati, S.; Rohani, R.; Basiri, M. E.; Abdar, M.; Yen, N. Y.; and Makarenkov, V. 2019. A Hybrid Latent Space Data Fusion Method for Multimodal Emotion Recognition. IEEE Access, 7: 172948--172964

2019
[28]

Poria, S.; Cambria, E.; Hussain, A.; and Huang, G.-B. 2015. Towards an intelligent framework for multimodal affective data analysis. Neural Networks, 63: 104--116

2015
[29]

Riquelme, C.; Puigcerver, J.; Mustafa, B.; Neumann, M.; Jenatton, R.; Susano Pinto, A.; Keysers, D.; and Houlsby, N. 2021. Scaling vision with sparse mixture of experts. Advances in Neural Information Processing Systems, 34: 8583--8595

2021
[30]

M.; Boureau, Y.-L.; and Weston, J

Roller, S.; Dinan, E.; Goyal, N.; Ju, D.; Williamson, M.; Liu, Y.; Xu, J.; Ott, M.; Smith, E. M.; Boureau, Y.-L.; and Weston, J. 2021. Recipes for Building an Open-Domain Chatbot. In Merlo, P.; Tiedemann, J.; and Tsarfaty, R., eds., Proceedings of the 16th Conference of the Eu...

2021
[31]

Romeo, L.; Cavallo, A.; Pepa, L.; Berthouze, N.; and Pontil, M. 2022. Multiple Instance Learning for Emotion Recognition Using Physiological Signals. IEEE Transactions on Affective Computing, 13: 389--407

2022
[32]

Shazeer, N.; Mirhoseini, A.; Maziarz, K.; Davis, A.; Le, Q.; Hinton, G.; and Dean, J. 2017. Outrageously large neural networks: The sparsely-gated mixture-of-experts layer. arXiv preprint arXiv:1701.06538

2017 arXiv
[33]

Su, H.; Liu, B.; Tao, J.; Dong, Y.; Huang, J.; Lian, Z.; and Song, L. 2020. An Improved Multimodal Dimension Emotion Recognition Based on Different Fusion Methods. In 2020 15th IEEE International Conference on Signal Processing (ICSP), volume 1, 257--261

2020
[34]

Sun, L.; Lian, Z.; Tao, J.; Liu, B.; and Niu, M. 2020. Multi-modal Continuous Dimensional Emotion Recognition Using Recurrent Neural Network and Self-Attention Mechanism. MuSe'20, 27–34. New York, NY, USA: Association for Computing Machinery. ISBN 9781450381574

2020
[35]

Sun, L.; Xu, M.; Lian, Z.; Liu, B.; Tao, J.; Wang, M.; and Cheng, Y. 2021. Multimodal Emotion Recognition and Sentiment Analysis via Attention Enhanced Recurrent Model. MuSe '21, 15–20. New York, NY, USA: Association for Computing Machinery. ISBN 9781450386784

2021
[36]

Tan, M.; and Le, Q. 2019. Efficientnet: Rethinking model scaling for convolutional neural networks. In International conference on machine learning, 6105--6114. PMLR

2019
[37]

Tang, H.; Liu, W.; Zheng, W.-L.; and Lu, B.-L. 2017. Multimodal Emotion Recognition Using Deep Neural Networks. 811--819

2017
[38]

Wang, G. 2024. Technical Research and Application Analysis of Multimodal Emotion Recognition. 2024 13th International Conference of Information and Communication Technology (ICTech), 109--113

2024
[39]

Williams, J.; Kleinegesse, S.; Comanescu, R.; and Radu, O. 2018. Recognizing Emotions in Video Using Multimodal DNN Feature Fusion. In Zadeh, A.; Liang, P. P.; Morency, L.-P.; Poria, S.; Cambria, E.; and Scherer, S., eds., Proceedings of Grand Challenge and Workshop on Human M...

2018
[40]

Wu, J.; Hu, X.; Wang, Y.; Pang, B.; and Soricut, R. 2023. Omni-SMoLA: Boosting Generalist Multimodal Models with Soft Mixture of Low-Rank Experts. 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 14205--14215

2023
[41]

Xu, W.; Jiang, H.; Liang, X.; and Liang, X. 2024. Leveraging Knowledge of Modality Experts for Incomplete Multimodal Learning. Proceedings of the 32nd ACM International Conference on Multimedia

2024
[42]

Ye, H.; and Xu, D. 2023. Taskexpert: Dynamically assembling multi-task representations with memorial mixture-of-experts. In Proceedings of the IEEE/CVF international conference on computer vision, 21828--21837

2023
[43]

Yu, W.; Xu, H.; Yuan, Z.; and Wu, J. 2021. Learning Modality-Specific Representations with Self-Supervised Multi-Task Learning for Multimodal Sentiment Analysis. Proceedings of the AAAI Conference on Artificial Intelligence, 35(12): 10790--10797

2021
[44]

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

Zadeh, A.; Liang, P. P.; Poria, S.; Vij, P.; Cambria, E.; and Morency, L.-P. 2018. Multi-attention recurrent network for human communication comprehension. In Proceedings of the Thirty-Second AAAI Conference on Artificial Intelligence and Thirtieth Innovative Applications of A...

2018
[45]

Zhao, Y.; and Chen, D. 2021. Expression EEG multimodal emotion recognition method based on the bidirectional LSTM and attention mechanism. Computational and Mathematical Methods in Medicine, 2021(1): 9967592

2021
[46]

Zheng, W.-L.; Liu, W.; Lu, Y.; Lu, B.-L.; and Cichocki, A. 2019. EmotionMeter: A Multimodal Framework for Recognizing Human Emotions. IEEE Transactions on Cybernetics, 49: 1110--1122

2019
[47]

Y.; Dai, A.; Chen, Z.; Le, Q.; and Laudon, J

Zhou, Y.; Lei, T.; Liu, H.; Du, N.; Huang, Y.; Zhao, V. Y.; Dai, A.; Chen, Z.; Le, Q.; and Laudon, J. 2022. Mixture-of-experts with expert choice routing. In Proceedings of the 36th International Conference on Neural Information Processing Systems, NIPS '22. Red Hook, NY, USA:...

2022
[48]

Ågren, W. 2022. The NT-Xent loss upper bound. arXiv:2205.03169

2022 arXiv
[49]

, " * 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...
[50]

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 6, 2026 · model on record in the stance chip above.

[1] [1]

Aktürk, T.; de Graaf, T.; Abra, Y.; Şahoğlu Göktaş, S.; Özkan, D.; Kula, A.; and Güntekin, B. 2021. Event-related EEG oscillatory responses elicited by dynamic facial expression. BioMedical Engineering OnLine, 20

work page 2021

[2] [2]

Alswaidan, N.; Menai, M.; Alswaidan, N.; and Menai, M. 2020. A survey of state-of-the-art approaches for emotion recognition in text. Knowledge and Information Systems, 62: 2937--2987

work page 2020

[3] [3]

ArunaGladys, A.; Vetriselvi, V.; ArunaGladys, A.; and Vetriselvi, V. 2023. Survey on multimodal approaches to emotion recognition. Neurocomputing, 556: 126693

work page 2023

[4] [4]

A.V., G.; T., M.; D., P.; and E., U. 2023. Multimodal Emotion Recognition with Deep Learning: Advancements, challenges, and future directions. Inf. Fusion, 105: 102218

work page 2023

[5] [5]

Binias, B.; Myszor, D.; Palus, H.; and Cyran, K. 2020. Prediction of Pilot's Reaction Time Based on EEG Signals. Frontiers in Neuroinformatics, 14

work page 2020

[6] [6]

Cheng, Y.; Li, Y.; He, J.; and Feng, R. 2024. Mixtures of experts for audio-visual learning. Advances in Neural Information Processing Systems, 37: 219--243

work page 2024

[7] [7]

Dai, D.; Dong, L.; Ma, S.; Zheng, B.; Sui, Z.; Chang, B.; and Wei, F. 2022. S table M o E : Stable Routing Strategy for Mixture of Experts. In Muresan, S.; Nakov, P.; and Villavicencio, A., eds., Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), 7085--7095. Dublin, Ireland: Association for Com...

work page 2022

[8] [8]

Fang, Y.; Huang, W.; Wan, G.; Su, K.; and Ye, M. 2025. EMOE: Modality-Specific Enhanced Dynamic Emotion Experts. In Proceedings of the Computer Vision and Pattern Recognition Conference, 14314--14324

work page 2025