pith. sign in

arxiv: 2502.21154 · v3 · submitted 2025-02-28 · 💻 cs.HC

Hypergraph Multi-Modal Learning for EEG-based Emotion Recognition in Conversation

Zijian Kang , Yueyang Li , Shengyu Gong , Weiming Zeng , Hongjie Yan , Lingbin Bian , Zhiguo Zhang , Wai Ting Siok
show 1 more author
Nizhuan Wang
This is my paper

Pith reviewed 2026-05-23 01:51 UTC · model grok-4.3

classification 💻 cs.HC
keywords EEG-based emotion recognitionhypergraph multi-modal learningconversation emotion recognitionmulti-modal fusionadaptive attentionphysiological signalsABEMA moduleAHFM module
0
0 comments X

The pith

Hyper-MML fuses EEG brain signals with audio and video through hypergraphs to recognize emotions in conversations more accurately than prior methods.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper develops Hyper-MML to recognize emotions during conversations by combining EEG physiological signals with audio and visual data. Existing methods mainly use semantic, audio, and visual inputs but struggle to integrate brain signals effectively. Two modules address this: ABEMA processes EEG across frequency bands with mutual-cross attention that adapts to individual subjects, while AHFM uses hypergraphs to link higher-order relationships among the modalities. Tests on the EAV and AFFEC datasets show the model outperforms current state-of-the-art approaches. The work targets applications in healthcare for conditions like autism and depression where patients may not express emotions verbally.

Core claim

Hyper-MML integrates EEG with audio and video information to capture complex emotional dynamics through an Adaptive Brain Encoder with Mutual-cross Attention module that extracts emotion-relevant features across frequency bands while adapting to subject-specific variations, combined with an Adaptive Hypergraph Fusion Module that actively models higher-order multi-modal relationships in emotion recognition in conversation.

What carries the argument

The Adaptive Hypergraph Fusion Module (AHFM) that models higher-order relationships among multi-modal signals, supported by the Adaptive Brain Encoder with Mutual-cross Attention (ABEMA) for EEG processing across frequency bands.

If this is right

  • The model significantly outperforms current state-of-the-art methods on the EAV and AFFEC datasets.
  • It can serve as an effective communication tool for healthcare professionals working with patients who struggle to express emotions.
  • The framework enables better engagement with individuals who have difficulty voicing their feelings.
  • It integrates physiological EEG signals into emotion recognition in conversation systems that previously relied mainly on audio and visual data.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The subject-adaptive attention in ABEMA could support personalized versions of the system for long-term use with specific patients.
  • Hypergraph fusion of physiological signals may apply to other multi-modal tasks such as mental health monitoring or human-computer interaction.
  • Real-time deployment in clinical settings could allow continuous emotion tracking during therapy sessions without relying on verbal reports.

Load-bearing premise

The ABEMA and AHFM modules capture emotion-relevant features and higher-order relationships in a way that generalizes beyond the tested subjects and the EAV and AFFEC datasets.

What would settle it

Evaluation on a new independent dataset with unseen subjects and conversation contexts where Hyper-MML fails to outperform state-of-the-art methods would falsify the performance claim.

Figures

Figures reproduced from arXiv: 2502.21154 by Hongjie Yan, Lingbin Bian, Nizhuan Wang, Shengyu Gong, Wai Ting Siok, Weiming Zeng, Yueyang Li, Zhiguo Zhang, Zijian Kang.

Figure 1
Figure 1. Figure 1: Illustration of emotion recognition in conversational text. The complete dialogue (a) shows coherent emotional [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Representation of Graphs and Hypergraphs. The upper left section displays a graph and its edges, while [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Overall framework of the Hyper-MML. In the context of EEG-based multimodal ERC, hypergraphs offer several compelling advantages over traditional graph-based approaches. The simultaneous interaction among EEG signals, audio features, and visual cues during emotional expression represents a natural higher-order relationship that cannot be effectively decomposed into pairwise connections without losing critic… view at source ↗
Figure 4
Figure 4. Figure 4: ABEMA for EEG embedding where WQD b , WKP b , WV P b , WQP b , WKD b , WV D b ∈ R C×dk are learnable parameter matrices, and dk is the hidden dimension. Then, we calculate bidirectional mutual-cross attention: A DP b = softmax QD b KP T √ b dk ! V P b , APD b = softmax QP b KDT √ b dk ! V D b , (8) Finally, we combine the outputs from both directions: Fb = A DP b + A PD b ∈ R B×dk , (9) Inter-band Mutual-C… view at source ↗
Figure 5
Figure 5. Figure 5: Confusion matrix on the EAV and AFFEC datasets. (a) EAV dataset shows confusion matrix for five [PITH_FULL_IMAGE:figures/full_fig_p015_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: T-SNE visualization of learned features on the EAV and AFFEC datasets. (a) EAV dataset visualization across [PITH_FULL_IMAGE:figures/full_fig_p015_6.png] view at source ↗
read the original abstract

Emotional Recognition in Conversation (ERC) is valuable for diagnosing health conditions such as autism and depression, and for understanding the emotions of individuals who struggle to express their feelings. Current ERC methods primarily rely on semantic, audio and visual data but face significant challenges in integrating physiological signals such as Electroencephalography (EEG). This research proposes Hypergraph Multi-Modal Learning (Hyper-MML), a novel framework for identifying emotions in conversation. Hyper-MML effectively integrates EEG with audio and video information to capture complex emotional dynamics. Firstly, we introduce an Adaptive Brain Encoder with Mutual-cross Attention (ABEMA) module for processing EEG signals. This module captures emotion-relevant features across different frequency bands and adapts to subject-specific variations through hierarchical mutual-cross attention mechanisms. Secondly, we propose an Adaptive Hypergraph Fusion Module (AHFM) to actively model the higher-order relationships among multi-modal signals in ERC. Experimental results on the EAV and AFFEC datasets demonstrate that our Hyper-MML model significantly outperforms current state-of-the-art methods. The proposed Hyper-MML can serve as an effective communication tool for healthcare professionals, enabling better engagement with patients who have difficulty expressing their emotions. The official implementation codes are available at https://github.com/NZWANG/Hyper-MML.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The manuscript proposes Hyper-MML, a multi-modal framework for emotion recognition in conversation that fuses EEG with audio and video modalities. It introduces the ABEMA module to process EEG via hierarchical mutual-cross attention across frequency bands while adapting to subject-specific variations, and the AHFM module to model higher-order relationships among modalities via hypergraphs. The central claim is that Hyper-MML significantly outperforms current SOTA methods on the EAV and AFFEC datasets, with code released at the provided GitHub link.

Significance. If the reported gains hold under subject-independent validation, the work would advance ERC by incorporating EEG signals, with potential utility in healthcare settings for patients with expression difficulties. The explicit release of implementation code is a clear strength for reproducibility.

major comments (2)
  1. [§4 (Experimental Setup) and Table 1/2] §4 (Experimental Setup) and Table 1/2: The claim that ABEMA 'adapts to subject-specific variations' and enables generalization is load-bearing for the central outperformance result, yet the manuscript does not report whether leave-one-subject-out (LOSO) cross-validation was used on EAV or AFFEC. Within-subject or random splits would allow subject-specific overfitting to explain the gains, undermining the adaptation mechanism.
  2. [§3.2 (AHFM) and ablation studies] §3.2 (AHFM) and ablation studies: No ablation isolates the contribution of the hypergraph fusion versus simpler concatenation or graph baselines; without this, it is unclear whether the reported improvements on multi-modal relationships are attributable to AHFM or to other factors such as increased model capacity.
minor comments (2)
  1. [Abstract] Abstract: The statement 'significantly outperforms current state-of-the-art methods' lacks any quantitative deltas, baseline names, or statistical test references.
  2. [§3.1] Notation in §3.1: The hierarchical mutual-cross attention equations use symbols (e.g., Q, K, V variants) that are not defined until later in the section, reducing readability.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive comments, which help clarify key aspects of our experimental validation and module contributions. We address each major point below with plans for revision.

read point-by-point responses

  1. Referee: [§4 (Experimental Setup) and Table 1/2] §4 (Experimental Setup) and Table 1/2: The claim that ABEMA 'adapts to subject-specific variations' and enables generalization is load-bearing for the central outperformance result, yet the manuscript does not report whether leave-one-subject-out (LOSO) cross-validation was used on EAV or AFFEC. Within-subject or random splits would allow subject-specific overfitting to explain the gains, undermining the adaptation mechanism.

    Authors: We agree that the validation procedure must be explicitly reported to substantiate the generalization claims for ABEMA. The revised manuscript will clearly state the cross-validation strategy used on EAV and AFFEC. In addition, we will perform and report new leave-one-subject-out (LOSO) experiments to confirm that the reported gains hold under subject-independent settings, directly addressing concerns about potential overfitting. revision: yes

  2. Referee: [§3.2 (AHFM) and ablation studies] §3.2 (AHFM) and ablation studies: No ablation isolates the contribution of the hypergraph fusion versus simpler concatenation or graph baselines; without this, it is unclear whether the reported improvements on multi-modal relationships are attributable to AHFM or to other factors such as increased model capacity.

    Authors: We acknowledge the value of isolating AHFM's contribution. The revised manuscript will include a new ablation study that directly compares the full AHFM hypergraph fusion against controlled baselines using concatenation and standard graph fusion, with model capacity matched across variants. This will clarify that performance gains stem from higher-order relational modeling rather than capacity differences. revision: yes

Circularity Check

0 steps flagged

No circularity: empirical claims rest on dataset experiments, not derivations or self-referential fits

full rationale

The paper introduces ABEMA and AHFM modules for EEG-audio-video fusion in ERC and supports its claims solely via performance numbers on the EAV and AFFEC datasets. No equations, derivations, parameter-fitting steps, or uniqueness theorems appear in the provided text. The central assertion (outperformance of SOTA) is therefore an empirical observation rather than a quantity that reduces by construction to its own inputs. No self-citation load-bearing steps, ansatzes, or renamings are present. This is the normal non-circular case for an applied ML methods paper whose validity hinges on external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract supplies no information on free parameters, axioms, or invented entities; cannot audit beyond the high-level module descriptions.

pith-pipeline@v0.9.0 · 5777 in / 1039 out tokens · 43047 ms · 2026-05-23T01:51:29.045041+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 2 Pith papers

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

  1. EEG-Based Multimodal Learning via Hyperbolic Mixture-of-Curvature Experts

    cs.LG 2026-04 unverdicted novelty 6.0

    EEG-MoCE uses learnable-curvature hyperbolic mixture-of-experts to capture hierarchical structures in multimodal EEG data and achieves state-of-the-art performance on emotion recognition, sleep staging, and cognitive ...

  2. EEG-Based Multimodal Learning via Hyperbolic Mixture-of-Curvature Experts

    cs.LG 2026-04 unverdicted novelty 6.0

    EEG-MoCE assigns each modality to a learnable-curvature hyperbolic expert and applies curvature-aware fusion to achieve state-of-the-art results on emotion recognition, sleep staging, and cognitive assessment benchmarks.

Reference graph

Works this paper leans on

40 extracted references · 40 canonical work pages · cited by 1 Pith paper · 1 internal anchor

  1. [1]

    Autism and Depression: clinical presentation, evaluation and treatment

    Amaia Hervás. Autism and Depression: clinical presentation, evaluation and treatment. Medicina (Argentina), 83(Suppl 2):37–42, 2023

    work page 2023

  2. [2]

    Emotion recognition in conversation: Research challenges, datasets, and recent advances

    Soujanya Poria, Navonil Majumder, Rada Mihalcea, and Eduard Hovy. Emotion recognition in conversation: Research challenges, datasets, and recent advances. IEEE access, 7:100943–100953, 2019

    work page 2019

  3. [3]

    MMGCN: Multimodal fusion via deep graph convolution network for emotion recognition in conversation

    Jingwen Hu, Yuchen Liu, Jinming Zhao, and Qin Jin. MMGCN: Multimodal fusion via deep graph convolution network for emotion recognition in conversation. arXiv preprint arXiv:2107.06779, 2021

    work page arXiv 2021

  4. [4]

    EEG based emotion recognition: A tutorial and review

    Xiang Li, Yazhou Zhang, Prayag Tiwari, Dawei Song, Bin Hu, Meihong Yang, Zhigang Zhao, Neeraj Kumar, and Pekka Marttinen. EEG based emotion recognition: A tutorial and review. ACM Computing Surveys, 55(4):1–57, 2022. 16 PRIME AI paper

    work page 2022

  5. [5]

    A tale of single-channel electroencephalogram: Devices, datasets, signal processing, applications, and future directions

    Yueyang Li, Weiming Zeng, Wenhao Dong, Di Han, Lei Chen, Hongyu Chen, Zijian Kang, Shengyu Gong, Hongjie Yan, Wai Ting Siok, et al. A tale of single-channel electroencephalogram: Devices, datasets, signal processing, applications, and future directions. IEEE Transactions on Instrumentation and Measurement, 2025

    work page 2025

  6. [6]

    Feature fusion based on mutual-cross-attention mechanism for eeg emotion recognition

    Yimin Zhao and Jin Gu. Feature fusion based on mutual-cross-attention mechanism for eeg emotion recognition. In International Conference on Medical Image Computing and Computer-Assisted Intervention, pages 276–285. Springer, 2024

    work page 2024

  7. [7]

    Eeg emotion copilot: Optimizing lightweight llms for emotional eeg interpretation with assisted medical record generation

    Hongyu Chen, Weiming Zeng, Chengcheng Chen, Luhui Cai, Fei Wang, Yuhu Shi, Lei Wang, Wei Zhang, Yueyang Li, Hongjie Yan, et al. Eeg emotion copilot: Optimizing lightweight llms for emotional eeg interpretation with assisted medical record generation. Neural Networks, page 107848, 2025

    work page 2025

  8. [8]

    Hypergcn: A new method for training graph convolutional networks on hypergraphs

    Naganand Yadati, Madhav Nimishakavi, Prateek Yadav, Vikram Nitin, Anand Louis, and Partha Talukdar. Hypergcn: A new method for training graph convolutional networks on hypergraphs. Advances in Neural Information Processing Systems, 32, 2019

    work page 2019

  9. [9]

    EA V: EEG-Audio-Video dataset for emotion recognition in conversational contexts.Scientific data, 11(1):1026, 2024

    Min-Ho Lee, Adai Shomanov, Balgyn Begim, Zhuldyz Kabidenova, Aruna Nyssanbay, Adnan Yazici, and Seong- Whan Lee. EA V: EEG-Audio-Video dataset for emotion recognition in conversational contexts.Scientific data, 11(1):1026, 2024

    work page 2024

  10. [10]

    Advancing face-to-face emotion communication: A multimodal dataset (affec)

    Meisam J Sekiavandi, Laurits Dixen, Jostein Fimland, Sree Keerthi Desu, Antonia-Bianca Zserai, Ye Sul Lee, Maria Barrett, and Paolo Burelli. Advancing face-to-face emotion communication: A multimodal dataset (affec). arXiv preprint arXiv:2504.18969, 2025

    work page arXiv 2025

  11. [11]

    Dialoguernn: An attentive rnn for emotion detection in conversations

    Navonil Majumder, Soujanya Poria, Devamanyu Hazarika, Rada Mihalcea, Alexander Gelbukh, and Erik Cambria. Dialoguernn: An attentive rnn for emotion detection in conversations. In Proceedings of the AAAI conference on artificial intelligence, volume 33, pages 6818–6825, 2019

    work page 2019

  12. [12]

    Dialoguegcn: A graph convolutional neural network for emotion recognition in conversation

    Deepanway Ghosal, Navonil Majumder, Soujanya Poria, Niyati Chhaya, and Alexander Gelbukh. Dialoguegcn: A graph convolutional neural network for emotion recognition in conversation. arXiv preprint arXiv:1908.11540, 2019

    work page arXiv 1908

  13. [13]

    Contrastive learning based modality-invariant feature acquisition for robust multimodal emotion recognition with missing modalities

    Rui Liu, Haolin Zuo, Zheng Lian, Björn W Schuller, and Haizhou Li. Contrastive learning based modality-invariant feature acquisition for robust multimodal emotion recognition with missing modalities. IEEE Transactions on Affective Computing, 15(4):1856–1873, 2024

    work page 2024

  14. [14]

    Mind the gap: Under- standing the modality gap in multi-modal contrastive representation learning

    Victor Weixin Liang, Yuhui Zhang, Yongchan Kwon, Serena Yeung, and James Y Zou. Mind the gap: Under- standing the modality gap in multi-modal contrastive representation learning. Advances in Neural Information Processing Systems, 35:17612–17625, 2022

    work page 2022

  15. [15]

    Eegnet: a compact convolutional neural network for eeg-based brain–computer interfaces

    Vernon J Lawhern, Amelia J Solon, Nicholas R Waytowich, Stephen M Gordon, Chou P Hung, and Brent J Lance. Eegnet: a compact convolutional neural network for eeg-based brain–computer interfaces. Journal of neural engineering, 15(5):056013, 2018

    work page 2018

  16. [16]

    Pgcn: Pyramidal graph convolutional network for eeg emotion recognition

    Ming Jin, Changde Du, Huiguang He, Ting Cai, and Jinpeng Li. Pgcn: Pyramidal graph convolutional network for eeg emotion recognition. IEEE Transactions on Multimedia, 26:9070–9082, 2024

    work page 2024

  17. [17]

    An efficient graph learning system for emotion recognition inspired by the cognitive prior graph of eeg brain network

    Cunbo Li, Tian Tang, Yue Pan, Lei Yang, Shuhan Zhang, Zhaojin Chen, Peiyang Li, Dongrui Gao, Huafu Chen, Fali Li, et al. An efficient graph learning system for emotion recognition inspired by the cognitive prior graph of eeg brain network. IEEE Transactions on Neural Networks and Learning Systems, 36(4):7130–7144, 2024

    work page 2024

  18. [18]

    Emotion recognition using multimodal residual lstm network

    Jiaxin Ma, Hao Tang, Wei-Long Zheng, and Bao-Liang Lu. Emotion recognition using multimodal residual lstm network. In Proceedings of the 27th ACM international conference on multimedia, pages 176–183, 2019

    work page 2019

  19. [19]

    Spatial–temporal recurrent neural network for emotion recognition

    Tong Zhang, Wenming Zheng, Zhen Cui, Yuan Zong, and Yang Li. Spatial–temporal recurrent neural network for emotion recognition. IEEE transactions on cybernetics, 49(3):839–847, 2018

    work page 2018

  20. [20]

    Tsception: Capturing temporal dynamics and spatial asymmetry from eeg for emotion recognition

    Yi Ding, Neethu Robinson, Su Zhang, Qiuhao Zeng, and Cuntai Guan. Tsception: Capturing temporal dynamics and spatial asymmetry from eeg for emotion recognition. IEEE Transactions on Affective Computing, 14(3):2238– 2250, 2022

    work page 2022

  21. [21]

    Accnet: Adaptive cross-frequency coupling graph attention for eeg emotion recognition

    Dongyuan Tian, Yucheng Wang, Peiliang Gong, Zhewen Xu, Zhenghua Chen, Xiaohui Wei, and Min Wu. Accnet: Adaptive cross-frequency coupling graph attention for eeg emotion recognition. Neural Networks, page 107853, 2025

    work page 2025

  22. [22]

    Eeg-based multimodal representation learning for emotion recognition

    Kang Yin, Hye-Bin Shin, Dan Li, and Seong-Whan Lee. Eeg-based multimodal representation learning for emotion recognition. In 2025 13th International Conference on Brain-Computer Interface (BCI), pages 1–4. IEEE, 2025

    work page 2025

  23. [23]

    A survey on hypergraph representation learning

    Alessia Antelmi, Gennaro Cordasco, Mirko Polato, Vittorio Scarano, Carmine Spagnuolo, and Dingqi Yang. A survey on hypergraph representation learning. ACM Computing Surveys, 56(1):1–38, 2023. 17 PRIME AI paper

    work page 2023

  24. [24]

    Random walks on hypergraphs with edge-dependent vertex weights

    Uthsav Chitra and Benjamin Raphael. Random walks on hypergraphs with edge-dependent vertex weights. In International Conference on Machine Learning, pages 1172–1181. PMLR, 2019

    work page 2019

  25. [25]

    Hgnn+: General hypergraph neural networks.IEEE Transactions on Pattern Analysis and Machine Intelligence, 45(3):3181–3199, 2022

    Yue Gao, Yifan Feng, Shuyi Ji, and Rongrong Ji. Hgnn+: General hypergraph neural networks.IEEE Transactions on Pattern Analysis and Machine Intelligence, 45(3):3181–3199, 2022

    work page 2022

  26. [26]

    Self-supervised hypergraph transformer for recommender systems

    Lianghao Xia, Chao Huang, and Chuxu Zhang. Self-supervised hypergraph transformer for recommender systems. In Proceedings of the 28th ACM SIGKDD conference on knowledge discovery and data mining, pages 2100–2109, 2022

    work page 2022

  27. [27]

    Exploiting spatial-temporal data for sleep stage classification via hypergraph learning

    Yuze Liu, Ziming Zhao, Tiehua Zhang, Kang Wang, Xin Chen, Xiaowei Huang, Jun Yin, and Zhishu Shen. Exploiting spatial-temporal data for sleep stage classification via hypergraph learning. In ICASSP 2024-2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pages 5430–5434. IEEE, 2024

    work page 2024

  28. [28]

    Exploring complex and heterogeneous correlations on hypergraph for the prediction of drug-target interactions

    Ding Ruan, Shuyi Ji, Chenggang Yan, Junjie Zhu, Xibin Zhao, Yuedong Yang, Yue Gao, Changqing Zou, and Qionghai Dai. Exploring complex and heterogeneous correlations on hypergraph for the prediction of drug-target interactions. Patterns, 2(12), 2021

    work page 2021

  29. [29]

    Multimodal fusion via hypergraph autoencoder and contrastive learning for emotion recognition in conversation

    Zijian Yi, Ziming Zhao, Zhishu Shen, and Tiehua Zhang. Multimodal fusion via hypergraph autoencoder and contrastive learning for emotion recognition in conversation. In Proceedings of the 32nd ACM international conference on multimedia, pages 4341–4348, 2024

    work page 2024

  30. [30]

    Neural-mcrl: Neural multimodal contrastive representation learning for eeg-based visual decoding

    Yueyang Li, Zijian Kang, Shengyu Gong, Wenhao Dong, Weiming Zeng, Hongjie Yan, Wai Ting Siok, and Nizhuan Wang. Neural-mcrl: Neural multimodal contrastive representation learning for eeg-based visual decoding. arXiv preprint arXiv:2412.17337, 2024

    work page arXiv 2024

  31. [31]

    iTransformer: Inverted Transformers Are Effective for Time Series Forecasting

    Yong Liu, Tengge Hu, Haoran Zhang, Haixu Wu, Shiyu Wang, Lintao Ma, and Mingsheng Long. itransformer: Inverted transformers are effective for time series forecasting. arXiv preprint arXiv:2310.06625, 2023

    work page internal anchor Pith review Pith/arXiv arXiv 2023

  32. [32]

    Opensmile: the munich versatile and fast open-source audio feature extractor

    Florian Eyben, Martin Wöllmer, and Björn Schuller. Opensmile: the munich versatile and fast open-source audio feature extractor. In Proceedings of the 18th ACM international conference on Multimedia, pages 1459–1462, 2010

    work page 2010

  33. [33]

    Learning deep global multi-scale and local attention features for facial expression recognition in the wild

    Zengqun Zhao, Qingshan Liu, and Shanmin Wang. Learning deep global multi-scale and local attention features for facial expression recognition in the wild. IEEE Transactions on Image Processing, 30:6544–6556, 2021

    work page 2021

  34. [34]

    Multivariate, multi-frequency and multimodal: Rethinking graph neural networks for emotion recognition in conversation

    Feiyu Chen, Jie Shao, Shuyuan Zhu, and Heng Tao Shen. Multivariate, multi-frequency and multimodal: Rethinking graph neural networks for emotion recognition in conversation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 10761–10770, 2023

    work page 2023

  35. [35]

    Mm-dfn: Multimodal dynamic fusion network for emotion recognition in conversations

    Dou Hu, Xiaolong Hou, Lingwei Wei, Lianxin Jiang, and Yang Mo. Mm-dfn: Multimodal dynamic fusion network for emotion recognition in conversations. In ICASSP 2022-2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pages 7037–7041. IEEE, 2022

    work page 2022

  36. [36]

    Graphmft: A graph network based multimodal fusion technique for emotion recognition in conversation

    Jiang Li, Xiaoping Wang, Guoqing Lv, and Zhigang Zeng. Graphmft: A graph network based multimodal fusion technique for emotion recognition in conversation. Neurocomputing, 550:126427, 2023

    work page 2023

  37. [37]

    A multi-view cnn with novel variance layer for motor imagery brain computer interface

    Ravikiran Mane, Neethu Robinson, A Prasad Vinod, Seong-Whan Lee, and Cuntai Guan. A multi-view cnn with novel variance layer for motor imagery brain computer interface. In 2020 42nd annual international conference of the IEEE engineering in medicine & biology society (EMBC), pages 2950–2953. Ieee, 2020

    work page 2020

  38. [38]

    Adversarial alignment and graph fusion via information bottleneck for multimodal emotion recognition in conversations

    Yuntao Shou, Tao Meng, Wei Ai, Fuchen Zhang, Nan Yin, and Keqin Li. Adversarial alignment and graph fusion via information bottleneck for multimodal emotion recognition in conversations. Information Fusion, 112:102590, 2024

    work page 2024

  39. [39]

    Decoding natural images from eeg for object recognition

    Yonghao Song, Bingchuan Liu, Xiang Li, Nanlin Shi, Yijun Wang, and Xiaorong Gao. Decoding natural images from eeg for object recognition. arXiv preprint arXiv:2308.13234, 2023

    work page arXiv 2023

  40. [40]

    Visual decoding and reconstruction via eeg embeddings with guided diffusion

    Dongyang Li, Chen Wei, Shiying Li, Jiachen Zou, Haoyang Qin, and Quanying Liu. Visual decoding and reconstruction via eeg embeddings with guided diffusion. arXiv preprint arXiv:2403.07721, 2024. 18

    work page arXiv 2024