Multi-Masked Querying Network for Robust Emotion Recognition from Incomplete Multi-Modal Physiological Signals

Geng-Xin Xu; Xiang Zuo; Ye Li

REVIEW 5 major objections 5 minor 26 references

Multi-Masked Querying Network for Robust Emotion Recognition from Incomplete Multi-Modal Physiological Signals

T0 review · 5 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read MMQ-Net claims that one masked transformer keeps emotion recognition above 90 percent accuracy even when 70 percent of the physiological signals are missing.

desk verdict The reconstruction loss never supervises missing-modality outputs, so the paper's central mechanism for missing data is untrained as written, despite a plausible architecture and useful question. read the letter →

arxiv 2507.20737 v1 pith:4NO2W6PO submitted 2025-07-28 cs.CV cs.AIcs.HC

Geng-Xin Xu , Xiang Zuo , Ye Li This is my paper · ORCID

classification cs.CVcs.AIcs.HC

keywords multi-modalemotionrecognitionphysiologicalsignalsmissingdatamaskedqueryingtransformerattentionmaskDEAPMAHNOB-HCIEEG

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

This paper tries to establish that a single transformer architecture, MMQ-Net, can recognize emotional valence and arousal from multi-modal physiological signals even when up to 70% of the signal modalities are missing and body-movement artifacts contaminate the rest. It reports that MMQ-Net outperforms seven comparison methods on DEAP and MAHNOB-HCI across all missing rates from 0 to 0.7, with the largest margins at the highest missing rates. The practical stake is real-time emotion monitoring for mental-health applications, where sensor dropouts and motion noise are unavoidable. If the results hold, the same design could serve as a front end for other incomplete multi-modal learning problems.

What carries the argument

The central object is the masked multi-head attention block in the Multi-Masked Querying Transformer, defined by a softmax attention with a mask matrix M derived from the modality-presence indicator a. M allows each query to attend only to available modalities plus the category and interference queries. The mechanism does three jobs at once: learnable modality queries occupy slots of missing data, category queries collect emotional-state evidence, and interference queries absorb noise; the three output streams are then supervised by respective losses. The mask is the load-bearing element: without it, missing queries would contaminate attention; with it, the model is claimed to learn which modalities are absent and reconstruct them from the present ones.

What would settle it

Measure reconstruction error on zeroed-out entries for which ground-truth features are known: if the model's loss on those entries does not drop below a mean-fill baseline, the claimed reconstruction of missing modalities is not what drives the accuracy.

Watch

Extended reading notes

Core claim

MMQ-Net's central claim is that missing-modality reconstruction and noise suppression can be unified in one masked multi-head attention block. The input sequence replaces absent modalities with learnable modality queries and appends a category query and an interference query; an attention mask built from the availability vector ensures that missing positions attend only to present data. The outputs split into reconstructed modality features, emotion features, and interference features, trained by three losses: feature-level reconstruction on present modalities, cross-entropy classification, and a mutual-information term that keeps emotion features label-relevant while decorrelating interference features from the label. On DEAP and MAHNOB-HCI, accuracy ranges from 90.06% to 99.85% depending on missing rate, exceeding the next best method by 2.37 to 6.76 percentage points.

Load-bearing premise

The reconstruction loss is computed only on modalities that are present, so the stated training procedure never directly supervises the model on values it is supposed to invent; the paper does not describe an extra protocol that hides available modalities and scores the imputed entries against ground truth.

Editorial extensions

If this is right

At a 0.7 missing rate, the method reports 90%+ accuracy on both datasets, implying it can be used when most modalities are absent.
The performance gap over the second-best method widens as missing rate increases, implying the benefits concentrate in data-scarce regimes.
Ablation dropping the reconstruction loss reduces accuracy by roughly 4 to 6 points at 30% missing, implying modality reconstruction contributes to classification.
Ablation dropping the interference loss also reduces accuracy, implying separation of noise from emotional features matters.
The same architecture and hyperparameters serve both DEAP and MAHNOB-HCI and both valence and arousal, implying the design transfers across datasets and targets.

Reading between the lines

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

We infer that the masked-query pattern is transferable to other partially observed multi-modal tasks, such as speech-emotion or human-activity recognition, since nothing in the design is physiological-signal-specific.
We note that because the reconstruction loss in Eq. (5) only uses available modalities, the attention mask may be acting mostly as a regularizer; a head-to-head comparison against oracle-supervised imputation would clarify.
We suggest a testable extension: hide a fixed fraction of present modalities during training with reconstruction targets, then report whether the held-out reconstruction error actually drops.
The near-99.85% accuracy at zero missing suggests the DEAP benchmark may be approaching saturation for this subject-dependent split, so future comparisons should include cross-subject protocols.

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.

Desk Editor's Note

read the letter

Colleague,

Bottom line: the paper's central mechanism for handling missing data is untrained as written. Eq. (5) computes the reconstruction loss only on available modalities: both the prediction and target are masked by a_i, so missing entries contribute zero. Nothing in the text says they artificially drop channels during training while keeping the original features as oracle targets. If that's the case, the modality queries are never supervised to output actual missing values, and the reported robustness gains must come from the category queries attending to available modalities, with LR acting as a regularizer rather than an imputer. The stress-test note is right.

What's genuinely new: the three-query combination (modality, category, interference) in one transformer, with the three-term loss, is an architectural tweak I haven't seen in the cited baselines. The accuracy tables are new measurements. The authors also correctly identify a real problem—missing modalities and artifact interference in wearable affect monitoring.

Credit where due: the ablation in Table 3 shows that removing LR hurts accuracy by 4-5 points, so the reconstruction loss is doing something. But that doesn't prove it reconstructs missing entries. The idea of using learnable query tokens to fill missing modalities is a reasonable extension of SMIL and TAE; it could work if the training protocol were specified and if the loss actually supervised imputation.

Soft spots, in order: (1) the reconstruction supervision gap is load-bearing; (2) no error bars, no subject-split description, no baseline provenance, so the 99.85% DEAP accuracy is uninterpretable; (3) Eq. (2) attention mask is under-specified—D+1[a^T,1,1] doesn't make sense as an attention mask without more explanation; (4) the mutual information loss in Eq. (7) is not operationalized; I(y;F_I|F_C) for continuous features is not obviously computable; (5) the title promises artifact interference, but no motion/artifact experiments are shown. These are real concerns, but none are insurmountable—they are fixable with a full protocol and honest experiments.

Who this is for: a reader working on missing-modality fusion in affective computing would find the architecture plausible and worth trying after the authors nail down the training protocol. Right now, the empirical claims outrun the evidence.

My take: I'd send this to peer review with a strong request for major revision—the question is meaningful and the architecture is plausible, but the paper as written doesn't support its central claim. I'd reject it in its current form, but I'd want to see the revision.

Referee Report

5 major / 5 minor

Summary. The paper proposes MMQ-Net, a transformer with three learnable query sets—modality queries, category queries, and interference queries—for emotion recognition from multimodal physiological signals with missing modalities and noise. It defines reconstruction, classification, and mutual-information losses, and reports accuracy on DEAP and MAHNOB-HCI across missing rates from 0.0 to 0.7, claiming consistent improvements over CCA, KCCA, DCCA, AE, SMIL, ShaSpe, and TAE. Ablation results show that removing either the reconstruction loss or the interference-reduction loss lowers accuracy. The central claims are that modality queries reconstruct missing data and that interference queries separate noise, but the manuscript provides no code, no detailed experimental protocol, and no direct validation of the interference mechanism.

Significance. The task is practically important, and the three-query decomposition is a plausible design direction. If the results were reproducible, a single architecture that robustly handles up to 70% missing physiological data would be a useful contribution. However, the reported evidence is not sufficient: the central reconstruction loss never supervises missing-modality outputs, the attention mask formula does not implement the claimed masking, the mutual-information loss is unspecified, and the experimental protocol lacks split definitions and error bars. The paper provides no code, machine-checked proofs, or parameter-free derivations, so the empirical tables are the sole support for the claims. The contribution may be salvageable in a substantially revised form, but as written it does not establish its claims.

major comments (5)

[§2.3, Eq. (5)] The reconstruction loss LR only compares masked predictions ai ⊙ F_hat_i^M with masked targets ai ⊙ F_i^M; for missing modalities ai,m = 0, both sides vanish, so the loss provides no gradient signal for the values output by modality queries for missing modalities. The experimental section describes missing rates only as a simulation of real-world conditions and never states that, during training, available channels are artificially dropped while the original features are retained as oracle targets. As written, the modality-query reconstruction mechanism is therefore untrained, and the reported robustness cannot be attributed to it.
[§2.3, Eq. (2)] The attention mask M = D + 1[a^T,1,1] is not a mask: it adds a positive constant to every position and leaves missing modalities with only the identity self-attention, rather than suppressing them. The text claims the mask ensures queries learn only from available modalities, but the formula does not implement this. A proper additive mask with -inf at missing tokens, or a clear derivation of why this formula masks missing positions, is needed.
[§2.4, Eq. (7)] The interference-reduction loss I(y; F_I | F_C) is written as a conditional mutual-information term, but the paper gives no estimator, no discretization, no variational bound, and no statement of how it is differentiated during training. Consequently the third objective is not well-defined as an implementable loss, and the ablation of LI cannot be interpreted without knowing what was actually computed.
[§3.2 and Tables 1-2] The empirical protocol is under-specified. There is no description of train/validation/test splits, no indication of whether evaluation is subject-dependent or subject-independent, no standard deviations or significance tests, and no implementation details for the seven baselines. Hyperparameters are said to be selected by cross-validation, but it is not stated that the numbers in the tables come from an independent test set. Without these details, the reported 90-99.85% accuracies and the claimed improvements over baselines are not verifiable.
[§3.4 and Abstract] The paper's second contribution, interference reduction, is never directly evaluated. The experiments only vary the modality missing rate; no synthetic artifacts, movement corruption, or real artifact-containing test set is used. The ablation that removes LI shows an accuracy drop, but that could be due to regularization rather than to interference separation. The interference-query mechanism therefore lacks empirical support.

minor comments (5)

[Fig. 2 caption] The caption contains a typo: 'Multi-model' should be 'Multi-modal'.
[§2.3, Eq. (2)] The notation 1[a^T,1,1] is ambiguous; the dimensions of the identity matrix and the resulting mask are not defined.
[§2.4, Eq. (5)] The denominator sum_i a_i^T 1 can be zero when all modalities are missing; the paper should state how this case is handled.
[§2.2] The list of frequency bands includes both alpha (8-10 Hz) and slow alpha (8-13 Hz); the overlap should be clarified.
[§3.2] No code release or data-processing script is mentioned, which limits reproducibility of the preprocessing and feature-extraction steps.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the reported gains are empirical outputs of trained losses, and the Eq. (5) masking issue is an evidence gap rather than a derivation that reduces to its inputs.

full rationale

The paper is an empirical method paper, and I find no load-bearing step in which a predicted quantity is identical by construction to a fitted input or to a self-cited prior result. The most serious weakness is in Sec. 2.4, Eq. (5): LR = (1 / sum_i a_i^T 1) * sum_i ||a_i ⊙ F_hat_i^M - a_i ⊙ F_i^M||^2. Because the availability vector a_i has entries 0 for missing modalities, both the prediction and the target are multiplied by zero there, so the loss is identically zero on exactly the entries that Sec. 2.3 says modality queries reconstruct. This undermines the paper's reconstruction narrative and is an undocumented-training-protocol concern, but it is not circular: the reported emotion-recognition accuracy is produced by training the classifier through LC (Eq. 6) on available features, and LR could act as a regularizer without ever supervising imputed values. The paper's self-citations (e.g., Refs. [8], [22], [24]) are background citations about prior emotion-recognition and EEG work; none is invoked as a uniqueness theorem, and none is the justification for the MMQ-Net architecture, so they are not load-bearing. The statement that hyper-parameters were selected based on cross-validation results is too underspecified to prove that the final tables were generated on the same folds used for tuning; even if that were so, it would be selection bias rather than a derivation that reduces to its own inputs. Therefore no circular step is established.

Assumptions & free parameters 3 free parameters · 5 assumptions · 0 invented entities

The central claim rests on a standard transformer backbone, DE/PSD feature assumptions, a simulated missingness protocol that is not described, and an unstated way to compute the mutual-information loss. The model introduces learnable query tokens, which are parameters rather than new physical or ontological entities. The main free parameters are the three loss weights and the network hyperparameters, all chosen by cross-validation without sensitivity analysis.

free parameters (3)

Loss weights lambda1, lambda2, lambda3 = lambda1=1, lambda2=1, lambda3=0.01
Chosen via cross-validation (Section 3.2); they trade reconstruction, classification, and interference reduction, and the paper gives no sensitivity analysis.
Learnable query tokens Q_M, Q_C, Q_I = Learned during training
Introduced in Eq. (1); these tokens replace missing features and act as readout and interference heads. Their initialization and dimension are not specified.
Network hyperparameters = 16-dim embedding, 4 heads, FFN 128, batch 1024, lr 6e-4, 5000 epochs
Selected based on cross-validation (Section 3.2); no grid or ablation is reported for these choices.

assumptions (5)

standard math Multi-head attention (Vaswani et al.) works as described and is the backbone of the querying transformer.
Used in Eq. (3); this is a standard background result.
domain assumption Differential entropy features assume a Gaussian distribution of band-limited EEG signals.
Stated in Section 2.2; if false, the DE features used for all modalities are biased.
domain assumption Simulated missing rates (0.1 to 0.7) represent real-world incomplete physiological signals.
Experiments drop modalities at some rate, but the paper never specifies how missingness is generated or whether it matches real sensor dropout (Section 3.2).
ad hoc to paper The conditional mutual information I(y; F_I | F_C) in Eq. (7) is computable and optimizable.
No estimator, variational bound, or approximation is provided; the loss as written cannot be directly implemented.
domain assumption Data labels (valence and arousal) in DEAP and MAHNOB-HCI are accurate and consistent across subjects.
Classification targets are taken as ground truth; no label noise modeling is discussed.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Multi-Masked Querying Network for Robust Emotion Recognition from Incomplete Multi-Modal Physiological Signals." pith.science (2026). https://pith.science/paper/4NO2W6PO

@misc{pith2026250720737,
  author       = {Pith},
  title        = {Pith review of: Multi-Masked Querying Network for Robust Emotion Recognition from Incomplete Multi-Modal Physiological Signals},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/4NO2W6PO}},
  note         = {Machine review of arXiv:2507.20737}
}

read the original abstract

Emotion recognition from physiological data is crucial for mental health assessment, yet it faces two significant challenges: incomplete multi-modal signals and interference from body movements and artifacts. This paper presents a novel Multi-Masked Querying Network (MMQ-Net) to address these issues by integrating multiple querying mechanisms into a unified framework. Specifically, it uses modality queries to reconstruct missing data from incomplete signals, category queries to focus on emotional state features, and interference queries to separate relevant information from noise. Extensive experiment results demonstrate the superior emotion recognition performance of MMQ-Net compared to existing approaches, particularly under high levels of data incompleteness.

Figures

Figures reproduced from arXiv: 2507.20737 by the authors.

**Figure 1.** Challenges in robust emotion recognition from incomplete multi-modal data. (a) Incomplete signals: Physiological signals with missing segments (dashed boxes), highlighting the incomplete learning problem. (b) Interference: Body movements and artifacts, complicating accurate emotion recognition. monitoring tools for clinical applications [26, 24, 16, 22]. In recent years, many researchers have explored emotion recog… view at source ↗

**Figure 2.** Flowchart of the MMQ-Net. Multi-model physiological signals are processed via encoders to extract features, which are then input into a Multi-Masked Querying Transformer to handle incomplete data and reduce interference. Hz bandpass filter is used to further clean the data, reducing the impact of measurement inaccuracies and environmental noise. Independent component analysis is then used to remove noise from signa… view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

26 extracted references · 25 canonical work pages

[1]

In: International conference on machine learning

Andrew, G., Arora, R., Bilmes, J., Livescu, K.: Deep canon ical correlation analysis. In: International conference on machine learning. pp. 1247 –1255. PMLR (2013)

work page 2013
[2]

Neural Networks 172, 106111 (2024)

Cheng, C., Liu, W., Fan, Z., Feng, L., Jia, Z.: A novel trans former autoencoder for multi-modal emotion recognition with incomplete data. Neural Networks 172, 106111 (2024)

work page 2024
[3]

In : Breakthroughs in statis- tics: methodology and distribution, pp

Hotelling, H.: Relations between two sets of variates. In : Breakthroughs in statis- tics: methodology and distribution, pp. 162–190. Springer (1992)

work page 1992
[4]

In: Proceedings of the Thirty-Third Internationa l Joint Conference on Ar- tiﬁcial Intelligence

Jia, Z., Zhao, F., Guo, Y., Chen, H., Jiang, T., Center, B.: Multi-level disentangling network for cross-subject emotion recognition based on mul timodal physiological signals. In: Proceedings of the Thirty-Third Internationa l Joint Conference on Ar- tiﬁcial Intelligence. pp. 3069–3077 (2024)

work page 2024
[5]

IEEE transactions on aﬀective computing 3(1), 18–31 (2011)

Koelstra, S., Muhl, C., Soleymani, M., Lee, J.S., Yazdani , A., Ebrahimi, T., Pun, T., Nijholt, A., Patras, I.: Deap: A database for emotion ana lysis; using physiolog- ical signals. IEEE transactions on aﬀective computing 3(1), 18–31 (2011)

work page 2011
[6]

In: ICASSP 20 19-2019 IEEE Inter- national Conference on Acoustics, Speech and Signal Proces sing (ICASSP)

Lee, H.C., Lin, C.Y., Hsu, P.C., Hsu, W.H.: Audio feature g eneration for missing modality problem in video action recognition. In: ICASSP 20 19-2019 IEEE Inter- national Conference on Acoustics, Speech and Signal Proces sing (ICASSP). pp. 3956–3960. IEEE (2019)

work page 2019
[7]

IEEE T ransactions on pattern analysis and machine intelligence 45(7), 8419–8432 (2023)

Lian, Z., Chen, L., Sun, L., Liu, B., Tao, J.: Gcnet: Graph c ompletion network for incomplete multimodal learning in conversation. IEEE T ransactions on pattern analysis and machine intelligence 45(7), 8419–8432 (2023)

work page 2023
[8]

IEEE Transactions on Consumer Electronics (2025) 10 G

Lin, Y., Xu, G.X., Liang, H., Wang, Y., Wan, F., Li, Y.: Brai n region knowledge based dual-stream transformer for eeg emotion recognition . IEEE Transactions on Consumer Electronics (2025) 10 G. Xu et al

work page 2025

Show all 26 references

[9]

IEEE Transactions on Aﬀective Com puting (2024)

Liu, R., Zuo, H., Lian, Z., Schuller, B.W., Li, H.: Contras tive learning based modality-invariant feature acquisition for robust multim odal emotion recognition with missing modalities. IEEE Transactions on Aﬀective Com puting (2024)

2024
[10]

In: Proceedings of the 31st ACM International Conference on Multimedia

Liu, Y., Jia, Z., Wang, H.: Emotionkd: a cross-modal know ledge distillation frame- work for emotion recognition based on physiological signal s. In: Proceedings of the 31st ACM International Conference on Multimedia. pp. 6122– 6131 (2023)

2023
[11]

In: International conferen ce on machine learning

Lopez-Paz, D., Sra, S., Smola, A., Ghahramani, Z., Schöl kopf, B.: Randomized nonlinear component analysis. In: International conferen ce on machine learning. pp. 1359–1367. PMLR (2014)

2014
[12]

MIT press (2014)

Luck, S.J.: An introduction to the event-related potent ial technique. MIT press (2014)

2014
[13]

In: Proceedings o f the AAAI Conference on Artiﬁcial Intelligence

Ma, M., Ren, J., Zhao, L., Tulyakov, S., Wu, C., Peng, X.: S MIL: Multimodal learning with severely missing modality. In: Proceedings o f the AAAI Conference on Artiﬁcial Intelligence. vol. 35, pp. 2302–2310 (2021)

2021
[14]

IEEE Journal of Biomedical and Health In formatics 26(8), 3607–3617 (2021)

Miao, M., Hu, W., Xu, B., Zhang, J., Rodrigues, J.J., De Al buquerque, V.H.C.: Automated cca-mwf algorithm for unsupervised identiﬁcati on and removal of eog artifacts from eeg. IEEE Journal of Biomedical and Health In formatics 26(8), 3607–3617 (2021)

2021
[15]

Trauma C are 4(1) (2024)

Miller, C.R., McDonald, J.E., Grau, P.P., Wetterneck, C .T.: Quality of life in posttraumatic stress disorder: The role of posttraumatic a nhedonia and depressive symptoms in a treatment-seeking community sample. Trauma C are 4(1) (2024)

2024
[16]

Information Fusion 112, 102590 (2024)

Shou, Y., Meng, T., Ai, W., Zhang, F., Yin, N., Li, K.: Adve rsarial alignment and graph fusion via information bottleneck for multimodal emotion recognition in conversations. Information Fusion 112, 102590 (2024)

2024
[17]

Sensors 18(7), 2074 (2018)

Shu, L., Xie, J., Yang, M., Li, Z., Li, Z., Liao, D., Xu, X., Yang, X.: A review of emotion recognition using physiological signals. Sensors 18(7), 2074 (2018)

2018
[18]

IEEE transactions on aﬀective computing 3(1), 42–55 (2011)

Soleymani, M., Lichtenauer, J., Pun, T., Pantic, M.: A mu ltimodal database for aﬀect recognition and implicit tagging. IEEE transactions on aﬀective computing 3(1), 42–55 (2011)

2011
[19]

Information Fusion 103, 102129 (2024)

Tang, J., Ma, Z., Gan, K., Zhang, J., Yin, Z.: Hierarchica l multimodal-fusion of physiological signals for emotion recognition with scenar io adaption and contrastive alignment. Information Fusion 103, 102129 (2024)

2024
[20]

Advances in ne ural information pro- cessing systems 30 (2017)

Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jon es, L., Gomez, A.N., Kaiser, Ł., Polosukhin, I.: Attention is all you need. Advances in ne ural information pro- cessing systems 30 (2017)

2017
[21]

In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recogni tion

Wang, H., Chen, Y., Ma, C., A very, J., Hull, L., Carneiro, G.: Multi-modal learning with missing modality via shared-speciﬁc feature modellin g. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recogni tion. pp. 15878– 15887 (2023)

2023
[22]

IEEE Trans- actions on Consumer Electronics (2025)

Wang, S., Zhou, T., Shen, Y., Li, Y., Huang, G., Hu, Y.: Gen erative ai enables eeg super-resolution via spatio-temporal adaptive diﬀusi on learning. IEEE Trans- actions on Consumer Electronics (2025)

2025
[23]

Current psychiatry r eports 23, 1–9 (2021)

Wilmer, M.T., Anderson, K., Reynolds, M.: Correlates of quality of life in anxiety disorders: review of recent research. Current psychiatry r eports 23, 1–9 (2021)

2021
[24]

In: Proceedings of the AAAI Conference on Artiﬁcial Intelligence

Zhang, L., Jin, L., Xu, G., Li, X., Xu, C., Wei, K., Liu, N., Liu, H.: Camel: capturing metaphorical alignment with context disentangling for mul timodal emotion recog- nition. In: Proceedings of the AAAI Conference on Artiﬁcial Intelligence. vol. 38, pp. 9341–9349 (2024)

2024
[25]

In: International Conference on M edical Image Computing and Computer-Assisted Intervention

Zhao, Y., Gu, J.: Feature fusion based on mutual-cross-a ttention mechanism for eeg emotion recognition. In: International Conference on M edical Image Computing and Computer-Assisted Intervention. pp. 276–285. Springe r (2024) MMQ-Net for Emotion Recognition from Incomplete M...

2024
[26]

IEEE Transactions on Aﬀe ctive Computing 15(3), 1358–1370 (2023)

Zhu, Q., Zheng, C., Zhang, Z., Shao, W., Zhang, D.: Dynami c conﬁdence-aware multi-modal emotion recognition. IEEE Transactions on Aﬀe ctive Computing 15(3), 1358–1370 (2023)

2023

Pith tools

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

[1] [1]

In: International conference on machine learning

Andrew, G., Arora, R., Bilmes, J., Livescu, K.: Deep canon ical correlation analysis. In: International conference on machine learning. pp. 1247 –1255. PMLR (2013)

work page 2013

[2] [2]

Neural Networks 172, 106111 (2024)

Cheng, C., Liu, W., Fan, Z., Feng, L., Jia, Z.: A novel trans former autoencoder for multi-modal emotion recognition with incomplete data. Neural Networks 172, 106111 (2024)

work page 2024

[3] [3]

In : Breakthroughs in statis- tics: methodology and distribution, pp

Hotelling, H.: Relations between two sets of variates. In : Breakthroughs in statis- tics: methodology and distribution, pp. 162–190. Springer (1992)

work page 1992

[4] [4]

In: Proceedings of the Thirty-Third Internationa l Joint Conference on Ar- tiﬁcial Intelligence

Jia, Z., Zhao, F., Guo, Y., Chen, H., Jiang, T., Center, B.: Multi-level disentangling network for cross-subject emotion recognition based on mul timodal physiological signals. In: Proceedings of the Thirty-Third Internationa l Joint Conference on Ar- tiﬁcial Intelligence. pp. 3069–3077 (2024)

work page 2024

[5] [5]

IEEE transactions on aﬀective computing 3(1), 18–31 (2011)

Koelstra, S., Muhl, C., Soleymani, M., Lee, J.S., Yazdani , A., Ebrahimi, T., Pun, T., Nijholt, A., Patras, I.: Deap: A database for emotion ana lysis; using physiolog- ical signals. IEEE transactions on aﬀective computing 3(1), 18–31 (2011)

work page 2011

[6] [6]

In: ICASSP 20 19-2019 IEEE Inter- national Conference on Acoustics, Speech and Signal Proces sing (ICASSP)

Lee, H.C., Lin, C.Y., Hsu, P.C., Hsu, W.H.: Audio feature g eneration for missing modality problem in video action recognition. In: ICASSP 20 19-2019 IEEE Inter- national Conference on Acoustics, Speech and Signal Proces sing (ICASSP). pp. 3956–3960. IEEE (2019)

work page 2019

[7] [7]

IEEE T ransactions on pattern analysis and machine intelligence 45(7), 8419–8432 (2023)

Lian, Z., Chen, L., Sun, L., Liu, B., Tao, J.: Gcnet: Graph c ompletion network for incomplete multimodal learning in conversation. IEEE T ransactions on pattern analysis and machine intelligence 45(7), 8419–8432 (2023)

work page 2023

[8] [8]

IEEE Transactions on Consumer Electronics (2025) 10 G

Lin, Y., Xu, G.X., Liang, H., Wang, Y., Wan, F., Li, Y.: Brai n region knowledge based dual-stream transformer for eeg emotion recognition . IEEE Transactions on Consumer Electronics (2025) 10 G. Xu et al

work page 2025