Pith. sign in

REVIEW 4 major objections 4 minor 40 references

Aligning Shared and Routed Experts for Cross-Subject EEG Generalization

T0 review · 4 major / 4 minor · reviewed 2026-08-03 · deepseek-v4-flash

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

Pith's one-line read The paper shows that cross-subject EEG generalization depends on a reducibility cost, and proposes a framework that pairs a shared expert with routed experts to outperform prior models on seven benchmarks.

desk verdict Plausible empirical combination, but the theory is asserted rather than proved and the evaluation lacks error bars; worth a serious look but not as-is. read the letter →

arxiv 2602.01728 v2 pith:2G5GTAAD submitted 2026-02-02 cs.CE

Zhi Zhang , Yan Liu , Zhejing Hu , Gong Chen , Sheng-hua Zhong , Changhong Jing , Shuqiang Wang , Jibin Wu
show 2 more authors
KC Tan Jiannong Cao
This is my paper · ORCID
classification cs.CE
keywords cross-subjectEEGdomaingeneralizationmixture-of-expertssharedexpertreducibilitycostalgorithmicalignmentmutual-guidedlearningdecoding
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 is trying to establish a design principle for cross-subject EEG models: whether to use one shared network or many specialized expert networks is not a matter of taste but depends on a reducibility cost, the sample-complexity price of approximating subject-specific functions by a single shared function. The authors prove a theorem (Theorem 2) that says a shared expert aligns better when that cost is low, and routed experts align better when it is high. They then build the Mutual-Guided Expert Collaboration (MGEC) framework, which runs a shared branch and a routed mixture-of-experts branch simultaneously and uses a mutual-guided reweighting to let each branch learn from the other on samples where the other is confident. On seven public EEG datasets, MGEC consistently outperforms state-of-the-art models, including EEG foundation models, which suggests the principle is more than theoretical. If correct, this gives practitioners a principled basis for choosing between invariant and subject-specific representations instead of guessing.

What carries the argument

The central object is the reducibility cost, defined through the sample-complexity bounds M(.,.) and the ratio alpha/beta in Theorem 2. The proof machinery is algorithmic alignment: a network aligns with a target function if its modules can be replaced by sub-functions with bounded sample complexity. The framework itself is carried by three mechanisms: (1) a shared expert trained with a joint embedding loss over augmented temporal neighbors to control transfer error; (2) a routed mixture-of-experts with prototype-based sparse routing and a subject-level entropy loss for specialization; (3) a mutual-guided reweighting that adjusts each branch's loss by the cross-branch gap exp(l_i^other - l_i

What would settle it

Train on a dataset where some samples are hard for both a shared model and a routed mixture-of-experts (e.g., synthetic data with overlapping class regions). If the mutual-guided loss makes training diverge or degrade compared to a fixed-weight ensemble, Assumption 4 is violated and the theory's guarantee breaks; if it stays stable, the mechanism is more forgiving than the assumption.

Watch

Extended reading notes

Core claim

The paper's central claim is that the divide between shared and routed architectures in cross-subject generalization is governed by a reducibility cost: the sample-complexity ratio of learning a branch-invariant function versus branch-specific ones. The authors formalize this in Theorem 2, which says a shared expert has better alignment when reducibility is cheap (small beta) and routed experts win when it is expensive (large alpha). They then instantiate the insight in MGEC, a framework that trains a shared branch and a routed mixture-of-experts branch jointly, with a mutual-guided reweighting that upweights samples where the better branch is confident and the other is not. Experiments on s

Load-bearing premise

The paper assumes that for any sample, at least one of the two branches is already close to correct; if both branches are wrong on hard or noisy samples, the mutual-guided reweighting will upweight the worse branch's error instead of correcting it.

Editorial extensions

If this is right

  • Practitioners can decide a priori whether to invest in a shared model, an expert mixture, or both, by estimating the reducibility cost on a validation set.
  • The mutual-guided reweighting is a plug-in regularizer that can be added to any shared/expert pair, not just EEG models, making it a general domain-generalization recipe.
  • The theory provides a principled explanation for why ensemble methods sometimes beat single models on heterogeneous subjects and sometimes hurt.
  • MGEC achieves consistent gains over strong backbones, including foundation models, across seven datasets spanning three EEG tasks.

Reading between the lines

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

  • A sharper test of the theory than the paper's synthetic experiment would be to measure per-subject reducibility cost on a real EEG dataset and verify that the optimal expert count tracks the predicted regime boundary.
  • The mutual-guided reweighting could be applied to any pair of models, not just EEG; this would make the framework a general recipe for domain generalization beyond brain signals.
  • If Assumption 4 fails on noisy samples, a confidence-gated reweighting that downweights samples where both branches disagree should be strictly more robust; the paper does not test this variant.
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 / 4 minor

Summary. The paper addresses cross-subject EEG generalization by combining a shared expert (for invariant, reducible functions) with routed mixture-of-experts (for subject-specific, irreducible functions), plus a mutual-guided reweighting scheme. The authors claim a theoretical justification based on algorithmic alignment: the choice between shared and routed experts depends on a 'reducibility cost' (Theorems 1-3), and they instantiate this in an MGEC framework. Experiments are reported on seven EEG benchmarks across auditory attention, motor imagery, and sleep staging, plus a synthetic domain-generalization experiment, with claims of consistent state-of-the-art performance.

Significance. The problem is important and the proposed architecture is a reasonable, well-motivated combination of two common paradigms. The breadth of the empirical evaluation — seven public datasets spanning three tasks and a synthetic verification — is a genuine strength. If the theoretical criterion were rigorously established, it would provide a principled design rule for when to use shared versus routed experts under distribution shift. However, the theory as stated is not currently load-bearing: Theorem 2 is an algebraic consequence of an assumed alignment formula that is never derived for the branching/MoE setting, and Theorem 3 is asserted without proof. The empirical claims also lack uncertainty quantification, and several gains are within 0.1-0.6 accuracy points. The contribution is therefore plausible but not yet established at the standard claimed in the paper.

major comments (4)
  1. [Section 2, Definition 2 / Theorem 2] Definition 1 defines alignment only for sequential compositions g = f_n ∘ ... ∘ f_1, with Align(N,g)=n·max_i M(f_i,N_i). Theorem 2 applies the same n·max_i formula to a branching target g(x)=Σ_j 1_{I_j}(h0(x))·h_j(x) and a routed-expert network R. A branching/gating function is not a composition of M+1 sequential sub-functions: the router and experts operate in parallel and are combined by a weighted sum. No proof is given that Align(R,g) = (M+1)·max(M(R0,h0), max_j M(R_j,h_j)), and the assumption M(R0,h0) ≤ max_j M(R_j,h_j) is simply asserted. Consequently, the 'reducibility cost' criterion is not derived from the stated framework; at present it is an unproved analogy. This is load-bearing for the paper's central theoretical claim.
  2. [Section 2, Theorem 3 / Assumption 4] Theorem 3 claims that, under Assumption 4, the total alignment error E[ε_S^(t)+ε_R^(t)] is non-increasing. No proof or training-dynamics model is provided. Assumption 4 — that for every sample at least one expert is ε-close to the target — is an unverified per-sample guarantee. The remark that it 'follows from Theorem 2' is not justified, since Theorem 2 concerns aggregate sample complexity and does not imply a per-sample bound. The mutual-guided losses in Section 3.3, Eqs. (10)-(11), are heuristic reweightings based on cross-entropy losses; if both branches are wrong on a hard/noisy sample, the weighting does not correct the worse branch. Without a proof or a direct empirical check of the learnability-gap premise, the claim that mutual-guided learning prevents over- and under-reduction is not supported.
  3. [Section 4.2, Tables 2-5] All comparison tables report point accuracies only; there are no standard deviations, confidence intervals, number of runs, or significance tests. Several reported improvements are very small: AVED is 53.9 vs 53.2 for the best baseline, and ISRUC-3 is 79.8 vs 79.4. Because the central empirical claim is that MGEC 'consistently outperforms state-of-the-art methods', these comparisons need repeated runs with paired significance tests, or at least error bars, before the claim can be evaluated. This is especially important given the small gains on some datasets.
  4. [Section 4.5, Synthetic Verification] The synthetic experiment is presented as empirical validation of the theoretical analysis, but it reports a single qualitative curve without error bars, and the mapping from the varying proportion of reducible components to the formal quantity α/β in Theorem 2 is not specified. As the only direct evidence for the theoretical criterion, the experiment is too weak to support the quantitative M+1 threshold in Theorem 2. Please provide repeated runs and a precise description of how the synthetic setup instantiates the reducibility cost.
minor comments (4)
  1. [Title/Abstract] The paper's metadata and the first abstract refer to 'Shared-Routed Expert Alignment (SREA)', while the full-text title and abstract refer to 'Mutual-Guided Expert Collaboration (MGEC)'. Please harmonize these names.
  2. [Table 5] In the last row, 'CbraMod Proposed' reads '76.579.8' — likely a missing column separator. Also, please clarify which backbone is used for the proposed method in the multi-task comparison and whether the improvement over CBraMod is statistically meaningful.
  3. [Section 4.4] The text states 'the six experts reduce to three groups', but Section 4.1 specifies M=5 experts. Please clarify this inconsistency.
  4. [Section 2] The term 'reducibility cost' is never formally defined; it is only the ratio α/β appearing in Theorem 2. A precise definition would help the reader connect the theoretical statement to the method description.

Circularity Check

2 steps flagged · score 6.0 of 10

Theorem 2's routed-vs-shared preference is an algebraic restatement of the very ratio used to define 'reducibility cost'; Assumption 4 is asserted to follow from an aggregate theorem it does not imply.

  1. self definitional [Section 2, Definition 2 and Theorem 2 (with Remark)]
    "Let α, β with 1≤α≤β satisfy α·max_j M(R_j, h_j, ε, δ) ≤ M(S_0, g, ε, δ) ≤ β·max_j M(R_j, h_j, ε, δ). Suppose the branching function is no harder to learn than the branch-dependent functions, i.e., M(R_0, h_0, ε, δ) ≤ max_j M(R_j, h_j, ε, δ). Then: (a) If α≥M+1, then Align(R, g, ε, δ) ≤ Align(S, g, ε, δ). (b) If β≤M+1, then Align(R, g, ε, δ) ≥ Align(S, g, ε, δ). Remark. The ratio α (or β) reflects the reducibility cost of branch-specific functions to a branch-invariant one."

    The theorem's conclusion is the same bracketing that defines α and β. With Align(S,g) taken as M(S0,g) and Align(R,g) taken as (M+1)·max_j M(R_j,h_j), clause (a) is just α≥M+1 ⇒ M(S0,g)≥(M+1)max_j M(...)=Align(R,g), and clause (b) is the reverse. No derivation extends the sequential alignment measure of Definition 1 to the parallel router-plus-experts architecture; the M+1 factor is assumed by construction. The 'reducibility cost' insight is therefore a restatement of the assumed ratio, not a predicted consequence.

  2. other [Section 2, Assumption 4 and Theorem 3]
    "Assumption 4. For any sample x, at least one expert produces a prediction close to the target function, i.e., min{∥S(x)−g(x)∥, ∥R(x)−g(x)∥} ≤ ε, where ε is a small constant. Remark. This follows from Theorem 2: at least one expert aligns well for each sample."

    The justification is not supplied: Theorem 2 is an aggregate sample-complexity alignment statement and does not imply a per-sample min-error guarantee for every x. The per-sample guarantee is exactly what Theorem 3 and the mutual-guided reweighting need; asserting it 'follows from Theorem 2' makes the theoretical support for the method rest on an unproved premise. This is an omitted derivation rather than a pure reduction, but it is load-bearing.

full rationale

The seven-benchmark empirical evaluation is independent, externally grounded, and not a fitted-input-called-prediction exercise, so the empirical claim is not circular. The theoretical core, however, is substantially definitional: Theorem 2's central criterion—'when this cost is low the shared expert aligns better; when it is high routed experts align better'—is obtained by defining α and β as the bracketing ratio and then reading off the same ratio as the conclusion, after implicitly assigning Align(R,g) an (M+1)·max form without a proof for the branching architecture. The 'reducibility cost' interpretation is thereby equivalent to the theorem's input. Separately, Assumption 4 is announced as following from Theorem 2, but Theorem 2 is aggregate and cannot yield a per-sample guarantee; Theorem 3 and the mutual-guided loss depend on this unsupported step. There is no significant self-citation chain or uniqueness-imported-from-authors issue, and the empirical benchmarks give the paper substantial independent content. Overall this is partial, construction-level circularity in the theoretical derivation, not a wholesale equivalence of the method to its inputs.

Assumptions & free parameters 4 free parameters · 6 assumptions · 2 invented entities

The central framework leans on a set of unverified distributional assumptions and on sample-complexity constructs adopted from algorithmic alignment. The invented 'reducibility cost' is a definitional ratio rather than an independently measured quantity. Free hyperparameters (M, K, ρ, loss weights) are fixed by hand without sensitivity analysis.

free parameters (4)
  • Number of routed experts M and active top-K = M=5, K=1
    Set by hand in Sec. 4.1 ('M=5 experts and activate one expert at a time'); no sensitivity analysis of M/K is reported.
  • Spatial-mask probability ρ = 0.1
    Used in shared-branch joint embedding loss δ(·,ρ); fixed at 10%, with temporal masking when electrodes <10 (Sec. 4.1).
  • Loss weighting coefficients = 1
    Sec. 4.1 states 'all coefficients are set to 1' and claims insensitivity without showing a sensitivity study.
  • Gate/prototype dimension d_r = not reported
    Dimensionality of Wz and prototype matrix D in Eq. 5 is not specified; needed for exact reproduction.
assumptions (6)
  • domain assumption Assumption 1: bounded density ratio C and transfer error η between train and test feature distributions
    Invoked for Theorem 1 generalization bounds; no verification on real EEG data.
  • domain assumption Assumption 2: an invariant function g_c exists that maps features to labels across train and test
    Underlies Theorem 1(a) and the shared-expert design.
  • domain assumption Assumption 3: a spurious function g_s fits training but fails on test
    Underlies Theorem 1(b) and the motivation for routing.
  • domain assumption Assumption 4: for every sample at least one of the shared or routed experts is ε-close to the target
    Needed for Theorem 3 and for the mutual reweighting to be beneficial rather than noise-amplifying.
  • standard math Algorithmic-alignment framework and sample-complexity measure M from Xu et al. 2020
    Borrowed without proof; all theorems are stated in terms of M.
  • domain assumption Branching structure of the target function (Definition 2) with partitions I_j and branch functions h_j
    Assumes subject heterogeneity can be modeled by a branch partition; the empirical datasets are never checked for this structure.
invented entities (2)
  • Reducibility cost (α/β)
    purpose: Theoretical quantity claimed to determine whether shared or routed experts align better; used to justify running both branches.
    Defined in Theorem 2 as the ratio of sample complexities; no estimator or independent falsifiable prediction is provided.
  • Mutual-guided learnability gap
    purpose: Sample-reweighting signal exp(ℓ_current − ℓ_guide) that transfers knowledge between shared and routed branches.
    Introduced in Eqs. 10–11; the claimed benefit is justified only by unproved Theorem 3.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Aligning Shared and Routed Experts for Cross-Subject EEG Generalization." pith.science (2026). https://pith.science/paper/2G5GTAAD

@misc{pith2026260201728,
  author       = {Pith},
  title        = {Pith review of: Aligning Shared and Routed Experts for Cross-Subject EEG Generalization},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/2G5GTAAD}},
  note         = {Machine review of arXiv:2602.01728}
}
read the original abstract

Cross-subject EEG generalization is challenging due to substantial heterogeneity across subjects. Existing methods typically learn either a shared subject-invariant model or multiple subject-specialized experts, but these two paradigms fail in complementary ways: the former may over-reduce subject-specific discriminative signals, while the latter may under-reduce transferable structure. We show that their suitability depends on the reducibility cost of branch-specific functions to branch-invariant ones, and we further provide a theory-to-method mapping that instantiates alignment principles in cross-subject EEG learning. Based on this insight, we propose Shared-Routed Expert Alignment (SREA), a collaborative framework that couples a shared expert for reducible invariant functions with routed experts for irreducible subject-specific functions. SREA trains the shared branch with joint embedding over augmented temporal neighbors, the routed branch with prototype-based sparse routing and expert specialization, and both branches with numerically stable mutual-guided reweighting based on cross-branch learnability gaps. Experiments on seven public EEG benchmarks across different tasks show that SREA consistently outperforms state-of-the-art methods and EEG foundation models.

Figures

Figures reproduced from arXiv: 2602.01728 by the authors.

Figure 1
Figure 1. The illustration of shared expert and routed expert behav [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. The framework of the proposed Mutual-Guided Expert Collaboration (MGEC) framework. [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Case studies of the proposed framework on the KUL dataset. The x-axis represents the subject index, and the y-axis represents the [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figures from the paper (1 more)
Figure 5
Figure 5. Figure 5: Experiments on synthetic verification under varying pro [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

40 extracted references

  1. [1]

    Aasm scoring manual updates for 2017,

    [Berryet al., 2017 ] Richard B Berry, Rita Brooks, Charlene Gamaldo, Susan M Harding, Robin M Lloyd, Stuart F Quan, Matthew T Troester, and Bradley V Vaughn. Aasm scoring manual updates for 2017,

    2017

  2. [6]

    The effect of head-related fil- tering and ear-specific decoding bias on auditory attention detection.Journal of Neural Engineering, 13(5):056014,

    [Daset al., 2016 ] Neetha Das, Wouter Biesmans, Alexander Bertrand, and Tom Francart. The effect of head-related fil- tering and ear-specific decoding bias on auditory attention detection.Journal of Neural Engineering, 13(5):056014,

    2016

  3. [9]

    Noise-robust cortical tracking of at- tended speech in real-world acoustic scenes.NeuroImage, 156:435–444,

    [Fuglsanget al., 2017 ] Søren Asp Fuglsang, Torsten Dau, and Jens Hjortkjær. Noise-robust cortical tracking of at- tended speech in real-world acoustic scenes.NeuroImage, 156:435–444,

    2017

  4. [10]

    Physiobank, physiotoolkit, and physionet: components of a new re- search resource for complex physiologic signals.Circu- lation, 101(23):e215–e220,

    [Goldbergeret al., 2000 ] Ary L Goldberger, Luis AN Ama- ral, Leon Glass, Jeffrey M Hausdorff, Plamen Ch Ivanov, Roger G Mark, Joseph E Mietus, George B Moody, Chung-Kang Peng, and H Eugene Stanley. Physiobank, physiotoolkit, and physionet: components of a new re- search resource for complex physiologic signals.Circu- lation, 101(23):e215–e220,

    2000

  5. [12]

    SalientSleepNet: Multimodal salient wave detection network for sleep stag- ing.IJCAI,

    [Jiaet al., 2021 ] Ziyu Jia, Youfang Lin, Jing Wang, Xuehui Wang, Peiyi Xie, and Yingbin Zhang. SalientSleepNet: Multimodal salient wave detection network for sleep stag- ing.IJCAI,

    2021

  6. [13]

    De- tecting the locus of auditory attention based on the spectro- spatial-temporal analysis of EEG.Journal of Neural En- gineering, 19(5):056035,

    [Jianget al., 2022 ] Yifan Jiang, Ning Chen, and Jing Jin. De- tecting the locus of auditory attention based on the spectro- spatial-temporal analysis of EEG.Journal of Neural En- gineering, 19(5):056035,

    2022

  7. [15]

    How EEG preprocessing shapes de- coding performance.Nature Communications Biology, 8(1):1039,

    [Kessleret al., 2025 ] Roman Kessler, Alexander Enge, and Michael A Skeide. How EEG preprocessing shapes de- coding performance.Nature Communications Biology, 8(1):1039,

    2025

  8. [16]

    ISRUC-Sleep: A comprehensive public dataset for sleep researchers

    [Khalighiet al., 2016 ] Sirvan Khalighi, Teresa Sousa, Jos´e Moutinho Santos, and Urbano Nunes. ISRUC-Sleep: A comprehensive public dataset for sleep researchers. Computer methods and programs in biomedicine, 124:180–192,

    2016

Show all 40 references
  1. [17]

    EEGNet: a compact convolutional neural network for EEG-based brain–computer interfaces

    [Lawhernet al., 2018 ] 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

  2. [20]

    Sparse mixture-of-experts are domain generalizable learn- ers

    [Liet al., 2023 ] Bo Li, Yifei Shen, Jingkang Yang, Yezhen Wang, Jiawei Ren, Tong Che, Jun Zhang, and Ziwei Liu. Sparse mixture-of-experts are domain generalizable learn- ers. InICLR,

    2023

  3. [23]

    DBPNet: Dual- branch parallel network with temporal-frequency fusion for auditory attention detection

    [Niet al., 2024 ] Qinke Ni, Hongyu Zhang, Cunhang Fan, Shengbing Pei, Chang Zhou, and Zhao Lv. DBPNet: Dual- branch parallel network with temporal-frequency fusion for auditory attention detection. InIJCAI,

    2024

  4. [24]

    U-time: A fully convolutional network for time series segmentation applied to sleep staging.NeurIPS, 32,

    [Perslevet al., 2019 ] Mathias Perslev, Michael Jensen, Sune Darkner, Poul Jørgen Jennum, and Christian Igel. U-time: A fully convolutional network for time series segmentation applied to sleep staging.NeurIPS, 32,

    2019

  5. [25]

    Ensemble deep learning and machine learn- ing: applications, opportunities, challenges, and future di- rections.Studies in Medical and Health Sciences, 1(2):18– 41,

    [Raneet al., 2024 ] Nitin Rane, Saurabh P Choudhary, and Jayesh Rane. Ensemble deep learning and machine learn- ing: applications, opportunities, challenges, and future di- rections.Studies in Medical and Health Sciences, 1(2):18– 41,

    2024

  6. [26]

    BCI2000: a general-purpose brain-computer inter- face (BCI) system.Transactions on Biomedical Engineer- ing, 51(6):1034–1043,

    [Schalket al., 2004 ] Gerwin Schalk, Dennis J McFarland, Thilo Hinterberger, Niels Birbaumer, and Jonathan R Wol- paw. BCI2000: a general-purpose brain-computer inter- face (BCI) system.Transactions on Biomedical Engineer- ing, 51(6):1034–1043,

    2004

  7. [28]

    Contrastive learning of subject-invariant EEG representations for cross-subject emotion recognition.Transactions on Affective Comput- ing, 14(3):2496–2511,

    [Shenet al., 2022 ] Xinke Shen, Xianggen Liu, Xin Hu, Dan Zhang, and Sen Song. Contrastive learning of subject-invariant EEG representations for cross-subject emotion recognition.Transactions on Affective Comput- ing, 14(3):2496–2511,

    2022

  8. [29]

    EEG conformer: Convolutional transformer for EEG decoding and visualization.Transactions on Neural Systems and Rehabilitation Engineering, 31:710–719,

    [Songet al., 2022 ] Yonghao Song, Qingqing Zheng, Bingchuan Liu, and Xiaorong Gao. EEG conformer: Convolutional transformer for EEG decoding and visualization.Transactions on Neural Systems and Rehabilitation Engineering, 31:710–719,

    2022

  9. [30]

    Towards recognizing spatial-temporal collaboration of EEG phase brain networks for emotion understanding

    [Sunet al., 2025 ] Jiangfeng Sun, Kaiwen Xue, Qika Lin, Yufei Qiao, Yifan Zhu, Zhonghong Ou, and Meina Song. Towards recognizing spatial-temporal collaboration of EEG phase brain networks for emotion understanding. In IJCAI, pages 3299–3307,

    2025

  10. [31]

    TinySleepNet: An efficient deep learning model for sleep stage scoring based on raw single-channel EEG

    [Supratak and Guo, 2020] Akara Supratak and Yike Guo. TinySleepNet: An efficient deep learning model for sleep stage scoring based on raw single-channel EEG. InEMBC, pages 641–644,

    2020

  11. [33]

    Review of the BCI competition IV.Frontiers in Neuroscience, 6:55,

    [Tangermannet al., 2012 ] Michael Tangermann, Klaus- Robert M ¨uller, Ad Aertsen, Niels Birbaumer, Christoph Braun, Clemens Brunner, Robert Leeb, Carsten Mehring, Kai J Miller, Gernot R M ¨uller-Putz, et al. Review of the BCI competition IV.Frontiers in Neuroscience, 6:55,

    2012

  12. [35]

    CBraMod: A criss-cross brain foundation model for EEG decoding

    [Wanget al., 2025 ] Jiquan Wang, Sha Zhao, Zhiling Luo, Yangxuan Zhou, Haiteng Jiang, Shijian Li, Tao Li, and Gang Pan. CBraMod: A criss-cross brain foundation model for EEG decoding. InICLR,

    2025

  13. [36]

    What can neural networks reason about? InICLR,

    [Xuet al., 2020 ] Keyulu Xu, Jingling Li, Mozhi Zhang, Si- mon S Du, Ken-ichi Kawarabayashi, and Stefanie Jegelka. What can neural networks reason about? InICLR,

    2020

  14. [37]

    DARNet: Dual attention refinement network with spatiotemporal construction for auditory attention detection.NeurIPS, 37:31688–31707,

    [Yanet al., 2024 ] Sheng Yan, Cunhang Fan, Hongyu Zhang, Xiaoke Yang, Jianhua Tao, and Zhao Lv. DARNet: Dual attention refinement network with spatiotemporal construction for auditory attention detection.NeurIPS, 37:31688–31707,

    2024

  15. [38]

    Rotboost: A technique for combining rota- tion forest and adaboost.Pattern Recognition Letters, 29(10):1524–1536,

    [Zhang and Zhang, 2008] Chun-Xia Zhang and Jiang-She Zhang. Rotboost: A technique for combining rota- tion forest and adaboost.Pattern Recognition Letters, 29(10):1524–1536,

    2008

  16. [40]

    CSBrain: A cross-scale spatiotemporal brain foundation model for EEG decoding.NeurIPS, 2025

    [Zhouet al., 2025b ] Yuchen Zhou, Jiamin Wu, Zichen Ren, Zhouheng Yao, Weiheng Lu, Kunyu Peng, Qihao Zheng, Chunfeng Song, Wanli Ouyang, and Chao Gou. CSBrain: A cross-scale spatiotemporal brain foundation model for EEG decoding.NeurIPS, 2025

    2025

  17. [1996]

    Low-latency auditory spatial attention detection based on spectro-spatial features from eeg

    [Caiet al., 2021 ] Siqi Cai, Pengcheng Sun, Tanja Schultz, and Haizhou Li. Low-latency auditory spatial attention detection based on spectro-spatial features from eeg. In EMBC, pages 5812–5815,

    2021

  18. [2000]

    Optimized bi-objective EEG chan- nel selection and cross-subject generalization with brain– computer interfaces.Transactions on Human-Machine Systems, 46(6):777–786,

    [Handiru and Prasad, 2016] Vikram Shenoy Handiru and Vinod A Prasad. Optimized bi-objective EEG chan- nel selection and cross-subject generalization with brain– computer interfaces.Transactions on Human-Machine Systems, 46(6):777–786,

    2016

  19. [2004]

    Deep learning with convolutional neural networks for EEG decoding and visualization.Human Brain Mapping, 38(11):5391–5420,

    [Schirrmeisteret al., 2017 ] Robin Tibor Schirrmeister, Jost Tobias Springenberg, Lukas Dominique Josef Fiederer, Martin Glasstetter, Katharina Eggensperger, Michael Tangermann, Frank Hutter, Wolfram Burgard, and Tonio Ball. Deep learning with convolutional neural networks for...

    2017

  20. [2007]

    Cross-subject emotion recognition using deep adaptation networks

    [Liet al., 2018 ] He Li, Yi-Ming Jin, Wei-Long Zheng, and Bao-Liang Lu. Cross-subject emotion recognition using deep adaptation networks. InInternational Conference on Neural Information Processing, pages 403–413,

    2018

  21. [2008]

    Evolutionary ensem- ble learning for eeg-based cross-subject emotion recog- nition.Journal of Biomedical and Health Informatics, 28(7):3872–3881,

    [Zhanget al., 2024 ] Hanzhong Zhang, Tienyu Zuo, Zhiyang Chen, Xin Wang, and Poly ZH Sun. Evolutionary ensem- ble learning for eeg-based cross-subject emotion recog- nition.Journal of Biomedical and Health Informatics, 28(7):3872–3881,

    2024

  22. [2012]

    EEG-based detection of the locus of auditory attention with convolutional neural net- works.Elife, 10:e56481,

    [Vandecappelleet al., 2021 ] Servaas Vandecappelle, Lucas Deckers, Neetha Das, Amir Hossein Ansari, Alexander Bertrand, and Tom Francart. EEG-based detection of the locus of auditory attention with convolutional neural net- works.Elife, 10:e56481,

    2021

  23. [2016]

    MSFNet: Multi-scale fusion network for brain-controlled speaker extraction

    [Fanet al., 2024 ] Cunhang Fan, Jingjing Zhang, Hongyu Zhang, Wang Xiang, Jianhua Tao, Xinhui Li, Jiangyan Yi, Dianbo Sui, and Zhao Lv. MSFNet: Multi-scale fusion network for brain-controlled speaker extraction. InMM, pages 1652–1661,

    2024

  24. [2017]

    Is limited participant diversity impeding EEG- based machine learning?NeurIPS,

    [Bomatter and Gouk, 2025] Philipp Bomatter and Henry Gouk. Is limited participant diversity impeding EEG- based machine learning?NeurIPS,

    2025

  25. [2018]

    Brain–computer communication: motiva- tion, aim, and impact of exploring a virtual apartment

    [Leebet al., 2007 ] Robert Leeb, Felix Lee, Claudia Kein- rath, Reinhold Scherer, Horst Bischof, and Gert Pfurtscheller. Brain–computer communication: motiva- tion, aim, and impact of exploring a virtual apartment. Transactions on Neural Systems and Rehabilitation Engi- neerin...

    2007

  26. [2019]

    Ensemble large language models: A sur- vey.Information, 16(8):688,

    [Mienye and Swart, 2025] Ibomoiye Domor Mienye and Theo G Swart. Ensemble large language models: A sur- vey.Information, 16(8):688,

    2025

  27. [2020]

    DeepSleepNet: A model for au- tomatic sleep stage scoring based on raw single-channel EEG.Transactions on Neural Systems and Rehabilitation Engineering, 25(11):1998–2008,

    [Suprataket al., 2017 ] Akara Supratak, Hao Dong, Chao Wu, and Yike Guo. DeepSleepNet: A model for au- tomatic sleep stage scoring based on raw single-channel EEG.Transactions on Neural Systems and Rehabilitation Engineering, 25(11):1998–2008,

    2017

  28. [2021]

    Neuromorphic sequential arena: A benchmark for neuromorphic temporal process- ing.IJCAI,

    [Chenet al., 2025 ] Xinyi Chen, Chenxiang Ma, Yujie Wu, Kay Chen Tan, and Jibin Wu. Neuromorphic sequential arena: A benchmark for neuromorphic temporal process- ing.IJCAI,

    2025

  29. [2022]

    Large brain model for learning generic repre- sentations with tremendous EEG data in BCI

    [Jianget al., 2024 ] Weibang Jiang, Liming Zhao, and Bao- liang Lu. Large brain model for learning generic repre- sentations with tremendous EEG data in BCI. InICLR,

    2024

  30. [2023]

    Reducing the subject variability of EEG signals with adversarial domain generalization

    [Maet al., 2019 ] Bo-Qun Ma, He Li, Wei-Long Zheng, and Bao-Liang Lu. Reducing the subject variability of EEG signals with adversarial domain generalization. In ICONIP, pages 30–42,

    2019

  31. [2024]

    Listen- Net: A lightweight spatio-temporal enhancement nested network for auditory attention detection.IJCAI,

    [Fanet al., 2025 ] Cunhang Fan, Xiaoke Yang, Hongyu Zhang, Ying Chen, Lu Li, Jian Zhou, and Zhao Lv. Listen- Net: A lightweight spatio-temporal enhancement nested network for auditory attention detection.IJCAI,

    2025

  32. [2025]

    Bagging predictors.Machine Learning, 24(2):123–140,

    [Breiman, 1996] Leo Breiman. Bagging predictors.Machine Learning, 24(2):123–140,

    1996

Pith tools

Reviewed August 3, 2026 · model on record in the stance chip above.