Toward the development of effective and efficient brain–computer interface (BCI) systems, precise decoding of brain activity measured by an electroencephalogram (EEG) is highly demanded. Traditional works classify EEG signals without considering the topological relationship among electrodes. However, neuroscience research has increasingly emphasized network patterns of brain dynamics. Thus, the Euclidean structure of electrodes might not adequately reflect the interaction between signals. To fill the gap, a novel deep learning (DL) framework based on the graph convolutional neural networks (GCNs) is presented to enhance the decoding performance of raw EEG signals during different types of motor imagery (MI) tasks while cooperating with the functional topological relationship of electrodes. Based on the absolute Pearson’s matrix of overall signals, the graph Laplacian of EEG electrodes is built up. The GCNs-Net constructed by graph convolutional layers learns the generalized features. The followed pooling layers reduce dimensionality, and the fully-connected (FC) softmax layer derives the final prediction. The introduced approach has been shown to converge for both personalized and groupwise predictions. It has achieved the highest averaged accuracy, 93.06% and 88.57% (PhysioNet dataset), 96.24% and 80.89% (high gamma dataset), at the subject and group level, respectively, compared with existing studies, which suggests adaptability and robustness to individual variability. Moreover, the performance is stably reproducible among repetitive experiments for cross-validation. The excellent performance of our method has shown that it is an important step toward better BCI approaches. To conclude, the GCNs-Net filters EEG signals based on the functional topological relationship, which manages to decode relevant features for brain MI. A DL library for EEG task classification including the code for this study is open source at https://github.com/SuperBruceJia/ EEG-DL for scientific research.

High precision and fast fault diagnosis is an important guarantee for the safe and reliable operation of machinery. In recent years, due to the strong recognition ability, data-driven fault diagnosis technology based on deep learning has attracted enormous attention. The fault diagnosis module proposed by many scholars has achieved excellent recognition results, but some of them are too complex to deploy in practice, due to the high costs. In this article, an efficient feature extraction method based on the convolutional neural networks (CNN) was proposed, and the high-precision fault diagnosis task was completed used a lightweight network. First, a 1-D CNN (1-D-CNN) is used to extract the multichannel features for the input original signals, to generate the feature maps which are highly consistent with the 2-D CNN (2-D-CNN). Second, a strategy of the lightweight model is proposed based on the principle of dynamic convolution and separable convolution. Third, a spatial attention mechanism (SAM) is used to adjust the weights of the output feature maps. Finally, the model completes the diagnosis task through the learned features. The performance of the model was verified under different operating conditions and noise environments. The experimental results demonstrate that the proposed method has excellent anti-noise ability and domain adaptability.

The quality of soybeans from different growing areas is different. It is common for low-quality soybeans to fake high-quality soybeans. This paper proposes a lightweight residual convolutional neural network (LRCNN) combined with an electronic nose (e-nose) to realize soybean quality traceability. Firstly, obtain soybean gas information from different growing areas through the e-nose. Then, according to the characteristics of e-nose detection data, the grouped heterogeneous kernel-based convolution (GHConv) is proposed, which effectively reduces the number of parameters through the combination of grouping and heterogeneous convolution. Finally, the LRCNN is proposed, which reduces the number of network parameters and avoids feature degradation, realizing the high-precision identification of soybean quality differences. In the multi-model comparison, the classification accuracy of the network is 98.37%, recall is 98.20%, and precision is 98.49%. The results show that the LRCNN combined with the e-nose can effectively identify the gas information of soybeans from different growing areas, providing a new method for soybean quality traceability.

The effective feature mining method is one of the key problems in the field of pattern recognition. Moreover, the lack of efficient feature extraction and selection methods has limited the application and development of electronic nose (e-nose) technology. In this study, a synergetic strategy for e-nose sensors data feature mining was proposed in combination with Support Vector Machine (SVM) to determine the beer olfactory information. First, twenty time-domain features and twenty frequency-domain features of e-nose sensors data were extracted to represent the olfactory characteristics of beer. Second, forty features were sorted with variable importance in projection (VIP) scores and forty subsets of multi-features with the best VIP score were generated. Finally, the classification models were established based on SVM, and the best parameter c and g of SVM models was calculated by Genetic Algorithm (GA). Furthermore, the classification performance of each class was evaluated by efficiency value (EFF) in different feature sets. The result indicates that GA-SVM model achieves good classification performance based on the #27 feature set with 81.67% and 96.67% in calibration set and testing set, respectively, and the EFF value is also the highest compared with other feature sets. In conclusion, it indicated that the analytical method can be used as a reliable tool for accurate identification of beer olfactory information.