Pith. sign in

REVIEW 4 major objections 6 minor 1 cited by

Robust Emotion Recognition via Bi-Level Self-Supervised Continual Learning

T0 review · 4 major / 6 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read SSOCL, a bi-level self-supervised continual learning framework with a pseudo-labeled dynamic memory buffer, lets an EEG emotion model adapt to unlabeled streaming subject data and generalize across subjects better than supervised and…

desk verdict Interesting new combination for EEG continual learning, but the central superiority claim is contradicted by its own Table 2 and the pseudo-label mechanism is never validated; not ready as written. read the letter →

arxiv 2505.10575 v2 pith:YYLHHG4R submitted 2025-05-13 cs.CV cs.LG

Adnan Ahmad , Bahareh Nakisa , Mohammad Naim Rastgoo This is my paper · ORCID
classification cs.CVcs.LG
keywords EEGemotionrecognitiononlinecontinuallearningself-supervisedmemoryreplaybufferpseudo-labelingtemporalcontrastivecross-subjectgeneralizationelectroencephalography
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

SSOCL is a self-supervised online continual learning framework that lets an EEG emotion recognition model keep learning from raw, unlabeled data as new subjects arrive in a stream, without falling into catastrophic forgetting. The paper's central contention is that this is possible using a bi-level architecture: an inner loop adapts a feature extractor to each incoming batch via temporal prediction and assigns pseudo-labels through k-means clustering with centroid-based cluster mapping, while a meta-loop trains the classifier on a dynamically curated memory buffer that retains only low-entropy, high-confidence pseudo-labeled samples. If the claim holds, emotion-recognition systems could be deployed in realistic settings where labeled data is scarce and emotions transition without clear boundaries, matching or exceeding methods that rely on labels. On two benchmark EEG datasets, the paper reports generalization accuracies of 71.78% on DEAP and 56.51% on AMIGOS, substantially above the best supervised baseline, and lower forgetting.

What carries the argument

The central machinery is the bi-level self-supervised loop. In the inner loop, a predictor network h maps each current embedding zt_n = f(xt_n) to a predicted next embedding, and a contrastive loss uses the true next embedding as the positive sample and all other predictions as negatives, forcing the feature extractor to preserve temporal dependencies without augmentation. K-means with K equal to the source class count partitions the adapted batch embeddings into clusters; a cluster-mapping module computes centroids for memory and current-batch clusters and assigns pseudo-labels by minimizing cosine distance between them. In the meta-loop, the model is trained with cross-entropy on the merged pseudo-labeled stream and memory data, and the memory enhancement module computes temperature-scaled softmax probabilities and their entropy, retaining only low-entropy, high-confidence samples in the fixed-size buffer. The memory buffer is therefore not just a replay store but a continuously refined pseudo-labeled set whose class boundaries are maintained across subject shifts.

What would settle it

Take a labeled EEG stream and run SSOCL with labels hidden; compare the pseudo-labels assigned by k-means and cluster mapping to the true labels for each batch, computing cluster purity or pseudo-label accuracy. If those scores are near chance, or if a deliberately single-class batch is split into K spurious clusters without the buffer correcting them, the central claim that the memory buffer stores reliable pseudo-labels collapses.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is that label-free online continual learning for EEG emotion recognition can be made to work by pairing two mechanisms: a self-supervised future-embedding prediction loss that adapts the feature extractor to each streamed batch without data augmentation, and a dynamic memory buffer whose contents and pseudo-labels are refined in a bi-level loop. The feature extractor is first adapted on the current batch by predicting the next embedding and applying a contrastive loss; k-means then partitions the batch into K clusters, and a cluster-mapping module aligns those clusters to memory clusters by cosine distance between centroids, assigning pseudo-labels. The meta-loop trains the model with cross-entropy on the merged stream-and-memory data, while a memory enhancement module uses temperature-scaled softmax entropy to keep only low-entropy samples in the fixed-size buffer. The paper presents this as a complete solution to cross-subject variability, noisy subjective labels, and blurred emotion boundaries in continuous physiological data streams, validated by experiments on DEAP and AMIGOS.

Load-bearing premise

The weakest assumption is that k-means clustering, run with K fixed to the source class count on every incoming batch, yields pseudo-labels that correspond to the true emotion categories in that batch, even when the batch contains only one class.

Editorial extensions

If this is right

  • If SSOCL is right, EEG emotion recognition models can keep adapting in real time to new users without any annotation effort, removing the need for subjective self-report labels that are noisy and costly.
  • The temporal-prediction objective is designed for low-signal-to-noise EEG and avoids the distortions that augmentation-based contrastive baselines suffer, which the ablation study supports.
  • The entropy-based memory curation should keep the buffer class-balanced and separable, as the UMAP visualizations show, making replay effective across evolving subjects.
  • The reported ForAcc values indicate the method forgets less than supervised and self-supervised baselines, so the model remains useful for previously seen subjects after learning new ones.

Reading between the lines

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

  • A natural extension is to measure pseudo-label accuracy directly against the true labels of the streamed subjects; the paper does not report this, and it would reveal how much of the accuracy gain comes from correct cluster assignments versus entropy-based filtering.
  • Because the same machinery is agnostic to signal type, it is plausible, though not shown, that temporal-prediction self-supervision plus entropy-buffered replay transfers to ECG, EMG, or wearable motion data with similar low-SNR, label-scarce streams.
  • Varying K from the fixed source-class count would test whether the cluster-mapping module truly resolves spurious partitions when a batch contains one or few classes; if accuracy is insensitive to K, the mapping and buffer curation are doing the corrective work the paper attributes to them.
  • The large run-to-run standard deviations, up to ±9.39 percentage points on AMIGOS, suggest the advantage over baselines may depend on subject order; reporting per-sequence significance tests would place the headline gain on firmer ground.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 6 minor

Summary. The paper proposes SSOCL, a bi-level self-supervised continual learning framework for EEG emotion recognition from streaming, unlabeled data. In an inner loop, a feature extractor is adapted to the current batch via a future-prediction contrastive loss; k-means clustering then partitions the batch into K clusters (K equal to the number of source classes), and a cluster-mapping module assigns pseudo-labels by comparing centroids with those of a memory buffer. In the meta-loop, an entropy-based memory enhancement module curates a fixed-size buffer, and the classifier is trained with cross-entropy on the merged pseudo-labeled current and memory data. The framework is evaluated in two transfer settings: PPB-EMO as source to DEAP, and PPB-EMO+DEAP as source to AMIGOS, reporting GenAcc of 71.78% (DEAP) and 56.51% (AMIGOS), with lower forgetting than baselines. The paper claims that SSOCL outperforms existing approaches in this label-free continual learning scenario.

Significance. If the central claims held, this would be a valuable contribution to affective computing: a truly label-free online continual learning method for EEG that handles cross-subject variability without distorting augmentations. The combination of temporal prediction for self-supervision, cluster-based pseudo-labeling, and entropy-based memory curation is original and practically motivated. However, the contribution is currently undermined by (i) an internal contradiction in the reported results on AMIGOS, (ii) a complete absence of validation of the pseudo-label mechanism on which the method depends, and (iii) an unusual contrastive loss with unaddressed degeneracy risks. The paper also has no code or machine-checked artifacts to substantiate reproducibility. The significance is therefore contingent on substantial further evidence, and is not yet established.

major comments (4)
  1. [Section 5.5, Table 2] The claim in Section 5.5 that 'SSOCL outperforms all baselines across all performance metrics' is directly contradicted by Table 2: on AMIGOS, EWC reaches AdapAcc 72.32% while SSOCL reaches AdapAcc 62.75%. This is not a minor reporting slip—it invalidates the stated superiority claim. The text must be corrected or the experiments re-examined, and any revised claim must account for this counterexample.
  2. [Sections 4.1.1, 4.1.2, 5.5] The pseudo-label assignment mechanism is the load-bearing component of the method, yet it is never validated. Section 4.1.1 admits that running k-means with K equal to the source class count on a batch containing a single class creates spurious partitions, and Section 4.1.2 states that the first batch is stored with arbitrary cluster-number pseudo-labels that become anchors for all later mappings via Eq. (4). There is no measurement of pseudo-label accuracy, cluster purity, or mapping reliability anywhere in the paper. Because Eq. (8) trains the classifier on these pseudo-labels and the memory buffer is curated using them, the reported gains cannot be attributed to the proposed label-free mechanism without evidence that the pseudo-labels correspond to the true emotion classes.
  3. [Section 4.1.1, Eq. (1)] Equation (1) defines a contrastive loss whose denominator sums over predicted embeddings \hat z_j, while the numerator is a similarity between a predicted embedding and the actual future embedding. This is not a standard InfoNCE formulation, and it is open to trivial solutions: if the predictor network outputs a near-constant vector for all inputs, the loss becomes uninformative because all predicted embeddings would be similar to each other. The paper offers no analysis, ablation, or theoretical justification for this design; the authors should either validate this loss on a simple diagnostic or replace it with a standard contrastive loss.
  4. [Table 3 vs Table 1] The ablation study in Table 3 reports SSOCL with AdapAcc 83.18%, GenAcc 72.96%, and ForAcc −9.92%, while Table 1 reports, for the same configuration on DEAP, AdapAcc 83.76%, GenAcc 71.78%, and ForAcc −15.05%. If these are the same experiment, the numbers should match; if they are from different runs or settings, that must be stated explicitly. The unexplained discrepancy reduces confidence in the reliability of the reported results.
minor comments (6)
  1. [Section 4.1.2, Eq. (4)] Equation (4) uses 'arg min_{i,j}' over a cosine-distance value, which yields a single pair rather than a mapping between all clusters of D_t and the memory; the intended assignment procedure should be described precisely (e.g., a one-to-one matching or per-cluster assignment).
  2. [Section 5.4] The statement '10 steps of training are considered in meta-loop' is ambiguous: clarify whether a step is one optimizer update or one pass over a mini-batch, and report the number of updates per time step and the total number of time steps.
  3. [Figure 1, caption] The caption uses 'current embeddings zt1' and 'future embeddings ˆzt2', while the main text indexes samples by n; unify the notation throughout the figure and text.
  4. [Tables 1 and 2] The standard deviations for SSOCL are substantially larger than those of the baselines (e.g., ±4.23 and ±6.60 on DEAP AdapAcc), yet no significance testing (e.g., paired tests across the five runs) is reported; this should be addressed to substantiate the performance differences.
  5. [Section 6.1] The temperature ablation is evaluated only through UMAP visualizations; the claim that T = 100 is optimal would be much stronger with a quantitative metric such as memory-buffer cluster purity or final accuracy per temperature.
  6. [References] References [5] and [6] appear to be the same paper (SimCLR); this duplication should be removed and the remaining citation fixed.

Circularity Check

0 steps flagged · score 0.0 of 10

No load-bearing circularity: pseudo-labels are an internal training signal, and GenAcc is measured against true labels.

full rationale

The central empirical claim is an accuracy result (DEAP GenAcc 71.78%, AMIGOS 56.51%) evaluated with ground-truth labels on a target test set comprising all subjects, not with the pseudo-labels used during training. The pseudo-label loop is a training-signal generator: Eq. 1 is a future-embedding contrastive objective; Eqs. 2-4 map k-means clusters to memory centroids; Eq. 5 merges pseudo-labeled batches; and Eq. 8 trains the classifier on those pseudo-labels. None of these equations produces the paper's headline evaluation quantity from its own input, so the reported generalization is not forced by construction. The fact that the model is trained on its own pseudo-labels is a standard pseudo-labeling procedure, not a circular derivation of the measured outcome. The only self-referential evidence is Figure 2, where the same trained model that assigned the pseudo-labels is used to embed memory samples and illustrate cluster separation; separability there is partly expected by construction, but that figure is illustrative and does not support the GenAcc comparison. The paper itself flags a genuine validity limitation in Section 4.1.1: when a batch contains a single class, k-means initialized with K 'will create unnecessary partitions, with only one cluster representing the true class,' and Section 4.1.2 stores the first batch with arbitrary cluster-number pseudo-labels. This is an unsupported-assumption concern about whether pseudo-labels match true emotion classes, not circularity: no equation reduces its output to its input, no fitted parameter is renamed as a prediction, and no load-bearing claim is justified by a self-citation chain.

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

The central mechanism depends on the clustering-based pseudo-labeling pipeline and the entropy-based memory selection. The main free parameters are K, T, and buffer size; these are chosen by hand rather than derived. The method introduces no new physical entities, forces, or dimensions.

free parameters (4)
  • Number of clusters K = 4
    Set equal to the number of emotion classes in the source dataset; k-means is initialized with K for every incoming batch, assuming each batch contains all classes (Section 4.1.1).
  • Temperature T for entropy scaling = 10 (setup) or 100 (ablation optimal)
    Controls the softmax sharpness in the memory enhancement module. Section 5.4 sets T=10, but Section 6.1 shows T=100 gives better separation; the value used for the main results is ambiguous.
  • Memory buffer size = 200
    Capacity of the replay buffer; set by hand and affects how many previous samples are retained (Section 5.4).
  • Contrastive temperature tau = not reported
    Temperature in the future-embedding contrastive loss (Eq. 1); its value is not given, but it scales the loss and is not a fixed physical constant.
assumptions (5)
  • domain assumption k-means with K clusters on self-supervised embeddings partitions each batch into true emotion classes
    Used in Section 4.1.1 to assign pseudo-labels; no cluster purity or pseudo-label accuracy is reported.
  • domain assumption Cosine centroid matching reliably maps new clusters to memory pseudo-labels
    Used in Section 4.1.2, Eq. 4; assumes high cosine similarity implies same class across streams.
  • domain assumption Low-entropy temperature-scaled predictions identify reliable samples
    Used in Section 4.2.1, Eq. 7; assumes model confidence correlates with label correctness under self-training.
  • domain assumption Temporal order within a batch provides a useful future-prediction pretext
    Used in Section 4.1.1, Eq. 1; assumes consecutive EEG segments are related enough that predicting the next embedding is a meaningful objective.
  • domain assumption The source-pretrained feature extractor transfers to target subjects
    The entire OCL setup starts from a model pretrained on PPB-EMO (and DEAP for AMIGOS); the method assumes this initialization is a good basis for adaptation.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Robust Emotion Recognition via Bi-Level Self-Supervised Continual Learning." pith.science (2026). https://pith.science/paper/YYLHHG4R

@misc{pith2026250510575,
  author       = {Pith},
  title        = {Pith review of: Robust Emotion Recognition via Bi-Level Self-Supervised Continual Learning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/YYLHHG4R}},
  note         = {Machine review of arXiv:2505.10575}
}
read the original abstract

Emotion recognition through physiological signals such as electroencephalogram (EEG) has become an essential aspect of affective computing and provides an objective way to capture human emotions. However, physiological data characterized by cross-subject variability and noisy labels hinder the performance of emotion recognition models. Existing domain adaptation and continual learning methods struggle to address these issues, especially under realistic conditions where data is continuously streamed and unlabeled. To overcome these limitations, we propose a novel bi-level self-supervised continual learning framework, SSOCL, based on a dynamic memory buffer. This bi-level architecture iteratively refines the dynamic buffer and pseudo-label assignments to effectively retain representative samples, enabling generalization from continuous, unlabeled physiological data streams for emotion recognition. The assigned pseudo-labels are subsequently leveraged for accurate emotion prediction. Key components of the framework, including a fast adaptation module and a cluster-mapping module, enable robust learning and effective handling of evolving data streams. Experimental validation on two mainstream EEG tasks demonstrates the framework's ability to adapt to continuous data streams while maintaining strong generalization across subjects, outperforming existing approaches.

Figures

Figures reproduced from arXiv: 2505.10575 by the authors.

Figure 1
Figure 1. SSOCL framework: Inner-loop: In a time step t, the feature extractor adapts to the current data Dt = x t in a self-supervised manner by using a self supervised module (SSM), which consists of a predictor network that takes current embeddings z t 1 and predicts future embeddings ˆz t 2 . After adaptation, k-means clustering partitions Zt into k clusters. To assign the pseudo labels, the cluster mapping module maps th… view at source ↗
Figure 2
Figure 2. The memory buffer obtained using the SSOCL approach at the end of training on [PITH_FULL_IMAGE:figures/full_fig_p016_2.png] view at source ↗
Figure 3
Figure 3. Visualization of the embedding spaces learned by [PITH_FULL_IMAGE:figures/full_fig_p017_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: UMAP visualizations of memory buffer embeddings for different temperature [PITH_FULL_IMAGE:figures/full_fig_p020_4.png]

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. Adaptive XAI in High Stakes Environments: Modeling Swift Trust with Multimodal Feedback in Human AI Teams

    cs.AI 2025-07 conditional novelty 4.0 of 10

    A conceptual framework uses implicit physiological signals to adapt AI explanations and support swift trust in high-stakes human-AI teams.

Reference graph

Works this paper leans on

49 extracted references · 45 canonical work pages · cited by 1 Pith paper

  1. [1]

    Robust federated learning under statistical heterogeneity via hessian-weighted aggregation

    Ahmad, A., Luo, W., Robles-Kelly, A., 2023. Robust federated learning under statistical heterogeneity via hessian-weighted aggregation. Ma- 21 chine Learning 112, 633–654. URL: https://api.semanticscholar. org/CorpusID:255427209

    work page 2023

  2. [2]

    Adaptive automation triggered by eeg-based mental workload index: A passive brain-computer interface application in real- istic air traffic control environment

    Aric´ o, P., Borghini, G., di Flumeri, G., Colosimo, A., Bonelli, S., Golfetti, A., Pozzi, S., Imbert, J.P., Granger, G., Benhac` ene, R., Ba- biloni, F., 2016a. Adaptive automation triggered by eeg-based mental workload index: A passive brain-computer interface application in real- istic air traffic control environment. Frontiers in Human Neuroscience

    work page

  3. [3]

    A passive brain-computer interface application for the mental workload assessment on professional air traffic controllers during realistic air traffic control tasks

    Aric´ o, P., Borghini, G., Flumeri, G.D., Colosimo, A., Pozzi, S., Babiloni, F., 2016b. A passive brain-computer interface application for the mental workload assessment on professional air traffic controllers during realistic air traffic control tasks. Progress in brain research 228, 295–328. URL: https://api.semanticscholar.org/CorpusID:205167096

    work page

  4. [4]

    Lur: An online learning model for eeg emotion recognition

    Cao, G., Yang, L., Zhang, Q., Xi, J., Tang, C., Tian, Y., 2023. Lur: An online learning model for eeg emotion recognition. 2023 IEEE International Conference on Bioinfor- matics and Biomedicine (BIBM) , 1038–1045URL: https://api. semanticscholar.org/CorpusID:267043725

    work page 2023

  5. [5]

    A simple framework for contrastive learning of visual representations, in: Pro- ceedings of the 37th International Conference on Machine Learning, pp

    Chen, T., Kornblith, S., Norouzi, M., Hinton, G., 2020a. A simple framework for contrastive learning of visual representations, in: Pro- ceedings of the 37th International Conference on Machine Learning, pp. 1597–1607

    work page

  6. [6]

    A sim- ple framework for contrastive learning of visual representations, in: ICML, pp

    Chen, T., Kornblith, S., Norouzi, M., Hinton, G.E., 2020b. A sim- ple framework for contrastive learning of visual representations, in: ICML, pp. 1597–1607. URL: http://proceedings.mlr.press/v119/ chen20j.html

    work page

  7. [7]

    A simple framework for contrastive learning of visual representations

    Chen, T., Kornblith, S., Norouzi, M., Hinton, G.E., 2020c. A simple framework for contrastive learning of visual representations. ArXiv abs/2002.05709. URL: https://api.semanticscholar.org/ CorpusID:211096730

    arXiv 2002

  8. [8]

    Amigos: A dataset for affect, personality and mood research on individuals and 22 groups

    Correa, J.A.M., Abadi, M.K., Sebe, N., Patras, I., 2017. Amigos: A dataset for affect, personality and mood research on individuals and 22 groups. IEEE Transactions on Affective Computing 12, 479–493. URL: https://api.semanticscholar.org/CorpusID:8743034

    work page 2017

Show all 49 references
  1. [9]

    Semisuper- vised autoencoders for speech emotion recognition

    Deng, J., Xu, X., Zhang, Z., Fr¨ uhholz, S., Schuller, B., 2018. Semisuper- vised autoencoders for speech emotion recognition. IEEE/ACM Trans- actions on Audio, Speech, and Language Processing 26, 31–43. URL: https://api.semanticscholar.org/CorpusID:2212886

    2018

  2. [10]

    URL: https://api.semanticscholar.org/CorpusID:6717657

  3. [11]

    Online continual decoding of streaming eeg signal with a balanced and informative memory buffer

    Duan, T., Wang, Z., Li, F., Doretto, G., Adjeroh, D., Yin, Y., Tao, C., 2024a. Online continual decoding of streaming eeg signal with a balanced and informative memory buffer. Neural networks : the official journal of the International Neural Network Society 176, 106338. URL: ...

  4. [12]

    Retain and adapt: Online sequential eeg classification with subject shift

    Duan, T., Wang, Z., Shen, L., Doretto, G., Adjeroh, D., Li, F., Tao, C., 2024b. Retain and adapt: Online sequential eeg classification with subject shift. IEEE Transactions on Artificial Intelligence URL: https: //api.semanticscholar.org/CorpusID:268984849

  5. [13]

    Self-supervised models are continual learners

    Fini, E., Costa, V., Alameda-Pineda, X., Ricci, E., Karteek, A., Mairal, J., 2021. Self-supervised models are continual learners. 2022 IEEE/CVF Conference on Computer Vision and Pattern Recogni- tion (CVPR) , 9611–9620URL: https://api.semanticscholar.org/ CorpusID:244954199

    2021

  6. [14]

    Mapping the signal- to-noise-ratios of cortical sources in magnetoencephalography and elec- troencephalography

    Goldenholz, D.M., Ahlfors, S.P., H¨ am¨ al¨ ainen, M.S., Sharon, D., Ishi- tobi, M., Vaina, L.M., Stufflebeam, S.M., 2009. Mapping the signal- to-noise-ratios of cortical sources in magnetoencephalography and elec- troencephalography. Human brain mapping 30, 1077–1086

    2009

  7. [15]

    Eeg in game user analysis: A framework for expertise classification during gameplay

    Hafeez, T., Saeed, S.M.U., Arsalan, A., Anwar, S.M., Ashraf, M.U., Alsubhi, K., 2021. Eeg in game user analysis: A framework for expertise classification during gameplay. PLoS ONE 16. URL: https://api. semanticscholar.org/CorpusID:231777428

    2021

  8. [16]

    Emotion recognition in human-computer interaction

    Hay, M.O., 2012. Emotion recognition in human-computer interaction. URL: https://api.semanticscholar.org/CorpusID:115023200

    2012

  9. [17]

    Hybrid spiking neural net- work for sleep electroencephalogram signals

    Jia, Z., Ji, J., Zhou, X., Zhou, Y., 2022. Hybrid spiking neural net- work for sleep electroencephalogram signals. Science China Information 23 Sciences 65. URL: https://api.semanticscholar.org/CorpusID: 247591578

    2022

  10. [18]

    Remonet: Reducing emotional label noise via multi-regularized self-supervision, in: ACM Multimedia

    Jiang, W.B., Lan, Y.T., Lu, B.L., 2024. Remonet: Reducing emotional label noise via multi-regularized self-supervision, in: ACM Multimedia. URL: https://api.semanticscholar.org/CorpusID:273645536

    2024

  11. [19]

    Variational deep embedding: An unsupervised and generative approach to clustering, in: International Joint Conference on Artificial Intelligence

    Jiang, Z., Zheng, Y., Tan, H., Tang, B., Zhou, H., 2016. Variational deep embedding: An unsupervised and generative approach to clustering, in: International Joint Conference on Artificial Intelligence. URL: https: //api.semanticscholar.org/CorpusID:2546662

    2016

  12. [20]

    Emo- tion recognition and artificial intelligence: A systematic review (2014–

    Khare, S.K., Blanes-Vidal, V., Nadimi, E.S., Acharya, U.R., 2024. Emo- tion recognition and artificial intelligence: A systematic review (2014–

    2024

  13. [21]

    Over- coming catastrophic forgetting in neural networks

    Kirkpatrick, J., Pascanu, R., Rabinowitz, N.C., Veness, J., Desjardins, G., Rusu, A.A., Milan, K., Quan, J., Ramalho, T., Grabska-Barwinska, A., Hassabis, D., Clopath, C., Kumaran, D., Hadsell, R., 2016. Over- coming catastrophic forgetting in neural networks. Proceedings of t...

    2016

  14. [22]

    Supervised contrastive learning

    Khosla, P., Teterwak, P., Wang, C., Sarna, A., Tian, Y., Isola, P., Maschinot, A., Liu, C., Krishnan, D., 2020. Supervised contrastive learning. Advances in neural information processing systems 33, 18661– 18673

    2020

  15. [23]

    Domain adaptation for eeg emotion recognition based on latent representation similarity

    Li, J., Qiu, S., Du, C., Wang, Y., He, H., 2020a. Domain adaptation for eeg emotion recognition based on latent representation similarity. IEEE Transactions on Cognitive and Developmental Systems 12, 344–353. URL: https://api.semanticscholar.org/CorpusID:208091556

  16. [24]

    Deap: A database for emo- tion analysis ;using physiological signals

    Koelstra, S., M¨ uhl, C., Soleymani, M., Lee, J.S., Yazdani, A., Ebrahimi, T., Pun, T., Nijholt, A., Patras, I., 2012. Deap: A database for emo- tion analysis ;using physiological signals. IEEE Transactions on Affec- tive Computing 3, 18–31. URL: https://api.semanticscholar.or...

    2012

  17. [25]

    A multimodal psychological, physiological and behavioural dataset for human emotions in driving tasks

    Li, W., Tan, R., Xing, Y., Li, G., Li, S., Zeng, G., Wang, P., Zhang, B., Su, X., Pi, D., Guo, G., Cao, D., 2022. A multimodal psychological, physiological and behavioural dataset for human emotions in driving tasks. Scientific Data 9. URL: https://api.semanticscholar.org/ Cor...

    2022

  18. [26]

    Multisource transfer learning for cross-subject eeg emotion recognition

    Li, J., Qiu, S., Shen, Y., Liu, C.L., He, H., 2020b. Multisource transfer learning for cross-subject eeg emotion recognition. IEEE Transactions 24 on Cybernetics 50, 3281–3293. URL: https://api.semanticscholar. org/CorpusID:89620182

  19. [27]

    Enhancing the reliability of out-of- distribution image detection in neural networks, in: International Con- ference on Learning Representations

    Liang, S., Li, Y., Srikant, R., 2018. Enhancing the reliability of out-of- distribution image detection in neural networks, in: International Con- ference on Learning Representations. URL: https://openreview.net/ forum?id=H1VGkIxRZ

    2018

  20. [28]

    A novel transfer- ability attention neural network model for eeg emotion recognition

    Li, Y., Fu, B., Li, F., Shi, G., Zheng, W., 2020c. A novel transfer- ability attention neural network model for eeg emotion recognition. ArXiv abs/2009.09585. URL: https://api.semanticscholar.org/ CorpusID:221819383

    2009 arXiv

  21. [29]

    Vbh-gnn: Variational bayesian heterogeneous graph neural networks for cross-subject emotion recognition, in: International Conference on Learning Representations

    Liu, C., qiu Zhou, X., Zhu, Z., Zhai, L., Jia, Z., Liu, Y., 2024. Vbh-gnn: Variational bayesian heterogeneous graph neural networks for cross-subject emotion recognition, in: International Conference on Learning Representations. URL: https://api.semanticscholar.org/ CorpusID:271745983

    2024

  22. [30]

    Cognitive Human-Machine Interfaces and Interactions for Avionics Systems

    Lim, Y.X., 2021. Cognitive Human-Machine Interfaces and Interactions for Avionics Systems. Phd thesis. RMIT University. URL: https:// doi.org/10.25439/rmt.27601791, doi:10.25439/rmt.27601791

    2021 doi

  23. [31]

    Driver emotion recog- nition with a hybrid attentional multimodal fusion framework

    Mou, L., Zhao, Y., Zhou, C., Nakisa, B., Rastgoo, M.N., Ma, L., Huang, T., Yin, B., Jain, R., Gao, W., 2023. Driver emotion recog- nition with a hybrid attentional multimodal fusion framework. IEEE Transactions on Affective Computing 14, 2970–2981. URL: https: //api.semanticsc...

    2023

  24. [32]

    Self-supervised learn- ing is more robust to dataset imbalance

    Liu, H., HaoChen, J.Z., Gaidon, A., Ma, T., 2021. Self-supervised learn- ing is more robust to dataset imbalance. ArXiv abs/2110.05025. URL: https://api.semanticscholar.org/CorpusID:238583191

    2021 arXiv

  25. [33]

    Evolutionary computation algorithms for feature selection of eeg-based emotion recognition using mobile sensors

    Nakisa, B., Rastgoo, M.N., Tjondronegoro, D., Chandran, V., 2018b. Evolutionary computation algorithms for feature selection of eeg-based emotion recognition using mobile sensors. Expert Syst. Appl. 93, 143–

  26. [34]

    Long short term memory hyperparameter optimization for a 25 neural network based emotion recognition framework

    Nakisa, B., Rastgoo, M.N., Rakotonirainy, A., Maire, F., Chandran, V., 2018a. Long short term memory hyperparameter optimization for a 25 neural network based emotion recognition framework. IEEE Access 6, 49325–49338. URL: https://api.semanticscholar.org/CorpusID: 52896590

  27. [35]

    Automatic driver stress level classification using multi- modal deep learning

    Rastgoo, M.N., Nakisa, B., Maire, F., Rakotonirainy, A., Chandran, V., 2019. Automatic driver stress level classification using multi- modal deep learning. Expert Syst. Appl. 138. URL: https://api. semanticscholar.org/CorpusID:198344222

    2019

  28. [36]

    Self-supervised ecg representation learning for emotion recognition

    Sarkar, P., Etemad, A., 2020. Self-supervised ecg representation learning for emotion recognition. IEEE Transactions on Affective Computing 13, 1541–1554. URL: https://api.semanticscholar.org/CorpusID: 211069320

    2020

  29. [37]

    Contrastive learning for unsupervised domain adaptation of time series

    Ozyurt, Y., Feuerriegel, S., Zhang, C., 2022. Contrastive learning for unsupervised domain adaptation of time series. ArXiv abs/2206.06243. URL: https://api.semanticscholar.org/CorpusID:249625545

    2022 arXiv

  30. [38]

    Vazquez-Rodriguez, J., Lefebvre, G., Cumin, J., Crowley, J.L.,

  31. [39]

    Self- supervised eeg emotion recognition models based on cnn

    Wang, X., Ma, Y., Cammon, J., Fang, F., Gao, Y., Zhang, Y., 2023. Self- supervised eeg emotion recognition models based on cnn. IEEE Transac- tions on Neural Systems and Rehabilitation Engineering 31, 1952–1962. URL: https://api.semanticscholar.org/CorpusID:257891756

    2023

  32. [40]

    Stochastic coherence over attention trajectory for continu- ous learning in video streams

    Tiezzi, M., Marullo, S., Faggi, L., Meloni, E., Betti, A., Melacci, S., 2022. Stochastic coherence over attention trajectory for continu- ous learning in video streams. ArXiv abs/2204.12193. URL: https: //api.semanticscholar.org/CorpusID:248392341

    2022 arXiv

  33. [41]

    Transformer-based self-supervised multimodal representation learning for wearable emotion recognition

    Wu, Y., Daoudi, M., Amad, A., 2023. Transformer-based self-supervised multimodal representation learning for wearable emotion recognition. IEEE Transactions on Affective Computing 15, 157–172. URL: https: //api.semanticscholar.org/CorpusID:257901125

    2023

  34. [42]

    Scale: Online self- supervised lifelong learning without prior knowledge

    Yu, X., Guo, Y., Gao, S., Simunic, T., 2022. Scale: Online self- supervised lifelong learning without prior knowledge. 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition Work- shops (CVPRW) , 2484–2495URL: https://api.semanticscholar. org/CorpusID:251765472

    2022

  35. [43]

    Individual emotion recog- nition and subgroup analysis from psychophysiological signals

    Zhang, L., Traue, H.C., Hazer-Rau, D., 2018. Individual emotion recog- nition and subgroup analysis from psychophysiological signals. Electrical Engineering eJournal

    2018

  36. [44]

    Im- proving task-free continual learning by distributionally robust memory 26 evolution, in: International Conference on Machine Learning

    Wang, Z., Shen, L., Fang, L., Suo, Q., Duan, T., Gao, M., 2022. Im- proving task-free continual learning by distributionally robust memory 26 evolution, in: International Conference on Machine Learning. URL: https://api.semanticscholar.org/CorpusID:250360832

    2022

  37. [45]

    Plug-and-play domain adaptation for cross-subject eeg-based emotion recognition, in: AAAI Conference on Artificial Intelligence

    Zhao, L.M., Yan, X., Lu, B.L., 2021. Plug-and-play domain adaptation for cross-subject eeg-based emotion recognition, in: AAAI Conference on Artificial Intelligence. URL: https://api.semanticscholar.org/ CorpusID:235306399. 27

    2021

  38. [48]

    Ganser: A self-supervised data augmentation framework for eeg-based emotion recognition

    Zhang, Z., hua Zhong, S., Liu, Y., 2021. Ganser: A self-supervised data augmentation framework for eeg-based emotion recognition. IEEE Transactions on Affective Computing 14, 2048–2063. URL: https:// api.semanticscholar.org/CorpusID:237431425

    2021

  39. [155]

    URL: https://api.semanticscholar.org/CorpusID:3203588

  40. [2022]

    2022 26th International Conference on Pattern Recogni- tion (ICPR) , 2605–2612URL: https://api.semanticscholar.org/ CorpusID:248084923

    Transformer-based self-supervised learning for emotion recog- nition. 2022 26th International Conference on Pattern Recogni- tion (ICPR) , 2605–2612URL: https://api.semanticscholar.org/ CorpusID:248084923

    2022

  41. [2023]

    Information Fusion 102, 102019

    and research recommendations. Information Fusion 102, 102019

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.