279 Hits in 6.9 sec

Transfer Learning for EEG-Based Brain-Computer Interfaces: A Review of Progress Made Since 2016 [article]

Dongrui Wu and Yifan Xu and Bao-Liang Lu
2020 arXiv   pre-print
Therefore, it is difficult to build a generic pattern recognition model in an EEG-based BCI system that is optimal for different subjects, during different sessions, for different devices and tasks.  ...  For each paradigm/application, we group the TL approaches into cross-subject/session, cross-device, and cross-task settings and review them separately.  ...  in adversarial attacks is different from the crosssubject TL setting in previous sections: in adversarial attacks, cross-subject means that the same machine learning model is used by all subjects, but  ... 
arXiv:2004.06286v4 fatcat:e32dqag5pvha7mzabrwead2hni

EEG-Based Emotion Recognition Using Regularized Graph Neural Networks [article]

Peixiang Zhong, Di Wang, Chunyan Miao
2020 arXiv   pre-print
In addition, we propose two regularizers, namely node-wise domain adversarial training (NodeDAT) and emotion-aware distribution learning (EmotionDL), to better handle cross-subject EEG variations and noisy  ...  Specifically, we model the inter-channel relations in EEG signals via an adjacency matrix in a graph neural network where the connection and sparseness of the adjacency matrix are inspired by neuroscience  ...  During optimization, our model aims to confuse the domain classifier by learning domain-invariant representations.  ... 
arXiv:1907.07835v4 fatcat:qh5dy2y7uzbc3ep2ozvr4zjac4

Deep Adversarial Domain Adaptation with Few-Shot Learning for Motor-Imagery Brain-Computer Interface

Chatrin Phunruangsakao, David Achanccaray, Mitsuhiro Hayashibe
2022 IEEE Access  
This study proposes the integration of deep domain adaptation with few-shot learning to address the challenge by leveraging the knowledge from multiple source subjects to enhance the performance of a single  ...  The domain discriminator was used to reduce domain drift, through adversarial training. The classifier predicted the user motor intention, based on EEG features.  ...  Owing to the intra-subject variability of EEG signals, the optimization of the classification model with cross-entropy loss may fail to generalize the samples from different sessions.  ... 
doi:10.1109/access.2022.3178100 fatcat:qgvxw4f64rcjlhkieyf7znnb24

Adversarial Representation Learning for Robust Patient-Independent Epileptic Seizure Detection [article]

Xiang Zhang, Lina Yao, Manqing Dong, Zhe Liu, Yu Zhang, Yong Li
2020 arXiv   pre-print
Methods: A complex deep neural network model is proposed to learn the pure seizure-specific representation from the raw non-invasive electroencephalography (EEG) signals through adversarial training.  ...  Furthermore, to enhance the explainability, we develop an attention mechanism to automatically learn the importance of each EEG channels in the seizure diagnosis procedure.  ...  [16] proposed a deep framework for the automatic detection of epileptic EEG by combining a stacked sparse autoencoder and a softmax classifier, which firstly learned the sparse and high-level representations  ... 
arXiv:1909.10868v2 fatcat:bsdfbhwkrna2vixujfnq46w7wq

Improving Cross-State and Cross-Subject Visual ERP-based BCI with Temporal Modeling and Adversarial Training

Ziyi Ni, Jiaming Xu, Yuwei Wu, Mengfan Li, Guizhi Xu, Bo Xu
2022 IEEE transactions on neural systems and rehabilitation engineering  
The results indicate that our hierarchical method can effectively model the long sequence EEG raw data, outperform the baselines on most conditions, including cross-state and cross-subject conditions.  ...  Finally, we show how deep learning-based methods with limited EEG data can improve ERP-based BCI with adversarial training. Our code will be released at  ...  Our model can be generally divided into three stages: (1) adversarial EEG generation, (2) temporal modeling, and (3) prediction and learning.  ... 
doi:10.1109/tnsre.2022.3150007 pmid:35133966 fatcat:d3p2zi5rwng75pq46xxamyka2a

Cross-Subject Domain Adaptation for Classifying Working Memory Load with Multi-Frame EEG Images [article]

Junfu Chen, Xiaoyi Jiang, Yang Chen, Bi Wang
2022 arXiv   pre-print
In this work, we propose a cross-subject deep adaptation model with spatial attention (CS-DASA) to generalize the workload classifications across subjects.  ...  First, the Subject-Shared module in CS-DASA receives multi-frame EEG image data from both source and target subjects and learns the common feature representations.  ...  And then, a joint optimization incorporating the goal of reducing the domain discrepancy and task loss of the source domain is conducted for the cross-subject domain adaptation.  ... 
arXiv:2106.06769v2 fatcat:ycbzhip5zvhvbnaj5nzwjgtchu

A Novel Neural Network Model based on Cerebral Hemispheric Asymmetry for EEG Emotion Recognition

Yang Li, Wenming Zheng, Zhen Cui, Tong Zhang, Yuan Zong
2018 Proceedings of the Twenty-Seventh International Joint Conference on Artificial Intelligence  
In this paper, we propose a novel neural network model, called bi-hemispheres domain adversarial neural network (BiDANN), for EEG emotion recognition.  ...  For further precisely predicting the class labels of testing data, we narrow the distribution shift between training and testing data by using a global and two local domain discriminators, which work adversarially  ...  Personalized (Subject-Independent) EEG Emotion Recognition In this experiment, we adopt a leave-one-subject-out cross validation strategy to evaluate the performance of our model, which is same with the  ... 
doi:10.24963/ijcai.2018/216 dblp:conf/ijcai/LiZCZZ18 fatcat:hjyfk7otmva55fwrm2d6oonfhu

Factorization Approach for Sparse Spatio-Temporal Brain-Computer Interface [article]

Byeong-Hoo Lee, Jeong-Hyun Cho, Byoung-Hee Kwon, Seong-Whan Lee
2022 arXiv   pre-print
From the results, we demonstrated that factorizing the EEG signal allows the model to extract rich and decisive features under sparse condition.  ...  To minimize the latent space shared by the class-common and class-specific features, the model is trained under orthogonal constraint.  ...  Adversarial Learning Adversarial learning trains models to solve the minmax optimization problem for the robustness of the models in several domains [49] - [53] .  ... 
arXiv:2206.08494v1 fatcat:qteotrwe2zgzfp6sv4cp2wbkxy

Single-Trial EEG Classification via Orthogonal Wavelet Decomposition-Based Feature Extraction

Feifei Qi, Wenlong Wang, Xiaofeng Xie, Zhenghui Gu, Zhu Liang Yu, Fei Wang, Yuanqing Li, Wei Wu
2021 Frontiers in Neuroscience  
Finally, sparse Bayesian learning with Gaussian prior is applied to the extracted power features, yielding an RVM classifier.  ...  Moreover, scalp weight maps of the spatial filters optimized by SEOWADE are more neurophysiologically meaningful.  ...  FUNDING This work was supported in part by the National Natural Science Foundation of China (No. 61906048, No. 61876063, and No. 61906019), Guangdong Basic and Applied Basic Research Foundation (No.  ... 
doi:10.3389/fnins.2021.715855 pmid:34720854 pmcid:PMC8548409 fatcat:sfm4fuugqzbbfdla5r7hib5pda

GANSER: A Self-supervised Data Augmentation Framework for EEG-based Emotion Recognition [article]

Zhi Zhang and Sheng-hua Zhong and Yan Liu
2021 arXiv   pre-print
especially deep learning models.  ...  As the first to combine adversarial training with self-supervised learning for EEG-based emotion recognition, the proposed framework can generate high-quality and high-diversity simulated EEG samples.  ...  They proposed a Wasserstein GAN-based framework to solve the domain shift problem by narrowing down the gap between the probability distribution of different subjects. Recently, Moon et al.  ... 
arXiv:2109.03124v2 fatcat:wwnxso7uwjdvzc3dwxuweb5bpq

Deep Learning in EEG: Advance of the Last Ten-Year Critical Period

Shu Gong, Kaibo Xing, Andrzej Cichocki, Junhua Li
2021 IEEE Transactions on Cognitive and Developmental Systems  
We first briefly mention the artifacts removal for EEG signal and then introduce deep learning models that have been utilized in EEG processing and classification.  ...  They are followed by the discussion, in which the pros and cons of deep learning are presented and future directions and challenges for deep learning in EEG are proposed.  ...  In contrast to the subject-specific models, the cross-subject model aims to have a general model for tolerating variance of subjects.  ... 
doi:10.1109/tcds.2021.3079712 fatcat:5rck4hvysfhe5o2tfjywytr5o4

Multi-View Spatial-Temporal Graph Convolutional Networks with Domain Generalization for Sleep Stage Classification

Ziyu Jia, Youfang Lin, Jing Wang, Xiaojun Ning, Yuanlai He, Ronghao Zhou, Yuhan Zhou, Li-Wei H Lehman
2021 IEEE transactions on neural systems and rehabilitation engineering  
and improve the generalization of deep neural networks is important. 3) Most deep learning methods ignore the interpretability of the model to the brain.  ...  Finally, domain generalization and MSTGCN are integrated into a unified framework to extract subject-invariant sleep features.  ...  Figure 10 illustrates that our model pays different attention to EEG channels in different sleep stages, which may caused by the EEG patterns of different sleep stages are different.  ... 
doi:10.1109/tnsre.2021.3110665 pmid:34487495 pmcid:PMC8556658 fatcat:kdpmv65nlfd6jiggqkicjsogvq

Table of Contents

2020 IEEE Signal Processing Letters  
Custom Domain Adaptation: A New Method for Cross-Subject, EEG-Based Cognitive Load Recognition .  ...  Boudraa 635 Amphibian Sounds Generating Network Based on Adversarial Learning . . . . . S. Park, M. Elhilali, D. K. Han, and H.  ...  Willett 1765 Cross-Informed Domain Adversarial Training for Noise-Robust Wake-Up Word Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  ... 
doi:10.1109/lsp.2020.3040840 fatcat:ezrfzwo6tjbkfhohq2tgec4m6y

An EEG Feature Extraction Method Based on Sparse Dictionary Self-Organizing Map for Event-Related Potential Recognition

Shang Feng, Haifeng Li, Lin Ma, Zhongliang Xu
2020 Algorithms  
Using sparse modeling to extract EEG signal features is a common approach.  ...  However, the features extracted by common sparse decomposition methods are only of analytical meaning, and cannot relate to actual EEG waveforms, especially event-related potential waveforms.  ...  The sparse modeling of the signal can be seen as solving the following optimization problem in Equation (2) .  ... 
doi:10.3390/a13100259 fatcat:uqi4xgdgr5afpfmqsbrof27twa

On the Metrics and Adaptation Methods for Domain Divergences of sEMG-based Gesture Recognition [article]

István Ketykó, Ferenc Kovács
2019 arXiv   pre-print
We assess our metric and the performance of our model in the context of sEMG/EMG-based gesture recognition under inter-session and inter-subject domain shifts.  ...  We propose a new metric to measure domain divergence and a new domain adaptation method for time-series classification.  ...  Domain-Adversarial Neural Network DANN) (Ganin et al., 2016) adversarially connects a binary domain classifier into the neural network directly exploiting the idea exhibited by Equation (19) .  ... 
arXiv:1912.08914v1 fatcat:ydbi5bzpqrhcziakafih3fbf6a
« Previous Showing results 1 — 15 out of 279 results