Current Issue

15 March 2021, Volume 3 Issue 1

Previous Issue    Next Issue

Review Intelligence

Parallel medicine: from warmness of medicare to medicine of smartness

Fei-Yue WANG

2021, 3(1):  1-9.  doi:10.11959/j.issn.2096-6652.202101

Asbtract ( 452 )   HTML ( 64)   PDF (399KB) ( 568 )   Knowledge map   

Figures and Tables | References | Related Articles | Metrics

How to combine intelligent technology with medical health and create more efficient intelligent medicine on the basis of traditional medicine and modern medicine is a historic task of great significance.Aiming at the complex problems of science, humanity and sociality of medicine, integrating complexity medicine, transdisciplinary medicine, and systems intelligence medicine, by using artificial societies for modeling and representation, computational experiments for verification and validation, parallel execution for management and control, the basic concepts, frames and processes of parallel medicine were proposed.As a systematic approach for smart medicine, the parallel medicine will produce medical big data from medical small data, and then generate medical smart data from medical big data.

Surveys and Prospectives

Intelligent service oriented fog radio access network:principles, technologies and challenges

Chenxi LIU, Binghong LIU, Xian ZHANG, Xinnan LONG, Mugen PENG

2021, 3(1):  10-17.  doi:10.11959/j.issn.2096-6652.202102

Asbtract ( 316 )   HTML ( 41)   PDF (2484KB) ( 471 )   Knowledge map   

Figures and Tables | References | Related Articles | Metrics

Big data and artificial intelligence has promoted the development of intelligent service.However, due to the high cost of data management and model training, as well as the increasing demands of users for latency and privacy, it is difficult to achieve the intelligent network based on the existing centralized network architecture.To address this issue, an artificial intelligence-based fog radio access network (AI-FRAN) architecture that supported distributed machine learning was proposed, and the fundamental principles that support the architecture were discussed.The key enabling techniques were identified that can realize the full utilization of communication resources, computing resources and cache resources in fog radio access network (F-RAN).Finally, the opportunities and challenges of AI-FRAN were discussed.

A survey on evolutionary ensemble learning algorithm

Yi HU, Boyang QU, Jing LIANG, Jie WANG, Yanli WANG

2021, 3(1):  18-35.  doi:10.11959/j.issn.2096-6652.202103

Asbtract ( 533 )   HTML ( 37)   PDF (478KB) ( 868 )   Knowledge map   

Figures and Tables | References | Related Articles | Metrics

Evolutionary ensemble learning integrates advantages of ensemble learning and evolutionary algorithm and is widely used in machine learning, data mining, and pattern recognition.Firstly, the theoretical basis, formation, and taxonomy are introduced.Secondly, according to the optimization task of evolutionary algorithm in ensemble learning, some representative researches on evolutionary ensemble learning field were analysed from the aspects of instance selection, feature selection, parameter optimization, structure optimization, and fusion strategy optimization, and the characteristics of different evolutionary ensemble learning methods were summarized.Finally, the pros and cons of the current researches on evolutionary ensemble learning were analysed, and research directions in the future work were given.

Special topic:emotional brain computer interface

A survey of affective brain-computer interface

Bao-Liang LU, Yaqian ZHANG, Wei-Long ZHENG

2021, 3(1):  36-48.  doi:10.11959/j.issn.2096-6652.202104

Asbtract ( 1090 )   HTML ( 209)   PDF (921KB) ( 1529 )   Knowledge map   

Figures and Tables | References | Related Articles | Metrics

An important research goal in emotion artificial intelligence is to make machines understand and recognize human emotions in real-time and facilitate human-computer interaction in a more natural and friendly way.Affective brain-computer interface (aBCI) is a type of BCI that can recognize and/or modulate human emotion.Thus, aBCI plays a critical role in promoting emotion artificial intelligence.The basic concepts and recent research development of aBCI were summarized, and the applications of aBCI in a wide range of domains were outlined.The roles that the aBCI can play in the development of artificial general intelligence and the challenges faced by the aBCI research community were discussed.

Emotional state decoding using EEG-based microstates of functional connectivity

Xinke SHEN, Yichao LI, Jin LIU, Sen SONG, Dan ZHANG

2021, 3(1):  49-58.  doi:10.11959/j.issn.2096-6652.202105

Asbtract ( 567 )   HTML ( 53)   PDF (1798KB) ( 838 )   Knowledge map   

Figures and Tables | References | Related Articles | Metrics

Emotional state decoding based on electroencephalography (EEG) usually regards individual emotion as a relatively static state and uses spectral power or inter-channel correlations of EEG as features.Based on recent advancement of dynamic functional connectivity analysis in the area of network neuroscience, a method called microstates of functional connectivity was designed and implemented, which clustered the inter-regional functional connectivity patterns of the brain under different emotional states to obtain representative microstates, and the temporal statistics, such as coverage and transition probability were extracted as features for emotional state decoding.Based on a widely used publicly available EEG dataset DEAP, new features in microstates of dynamic functional connectivity analysis achieved regression mean squared errors of 3.87±0.28 and 3.25±0.30 on valence and arousal respectively, which were better than those using traditional spectral power features, 4.07±0.30 (p=0.005) and 3.41±0.31 (p=0.064).The results demonstrate the feasibility of emotional state decoding based on microstates of functional connectivity and provide deeper insight into understanding the neural mechanisms of emotion.

Cross-subject emotional EEG recognition based on multi-source domain adaptation

Hanbing GAO, Chi ZHANG, Mingyan JIN, Yang XIAO, Fengyu CONG

2021, 3(1):  59-64.  doi:10.11959/j.issn.2096-6652.202106

Asbtract ( 457 )   HTML ( 48)   PDF (1808KB) ( 681 )   Knowledge map   

Figures and Tables | References | Related Articles | Metrics

During the recognition of emotion based on electroencephalography (EEG) signals, traditional machine learning and deep learning methods cannot establish a general classification and detection model for EEG data due to the differences in EEG signals among subjects.Each subject was treated as an independent domain, a multi-source cross-subject emotional EEG recognition model was established, and the model was verified on the public dataset.The evaluation results show that compared with single-source domain model, the proposed model has better cross-subject feature extraction and classification capabilities.

Emotion recognition based on brain and machine collaborative intelligence

Dongjun LIU, Yuhan WANG, Wenfen LING, Yong PENG, Wanzeng KONG

2021, 3(1):  65-75.  doi:10.11959/j.issn.2096-6652.202107

Asbtract ( 610 )   HTML ( 56)   PDF (2762KB) ( 1232 )   Knowledge map   

Figures and Tables | References | Related Articles | Metrics

Emotion recognition is a direct and effective mode of emotion recognition.Machine learning relies on the formal representation of image expressions, lacks the cognitive representation ability of the brain, and has poor recognition performance on small sample data sets or complex expression (camouflage) data sets.To this end, the formal representation of machine artificial intelligence was combined with the emotional cognitive ability of human brain general intelligence, and a brain-machine collaborative intelligence emotion recognition method was proposed.Firstly, electroencephalogram (EEG) emotional features were extracted from EEG to obtain the brain’s cognitive representation of emotions.Secondly, the visual features of the image were extracted from the emotional image to obtain the machine’s formal representation of the emotion.In order to enhance the generalization ability of the machine model, the transfer adaptation between samples was introduced in the feature learning.After obtaining image visual features and EEG emotional features, the random forest regression model was trained to obtain the brain-machine mapping relationship between image visual features and EEG emotional features.The visual features of the test image were generated through the brain-machine mapping relationship to generate virtual EEG emotional features, and then the virtual EEG emotional features and image visual features were fused for emotion recognition.This method has been verified on the Chinese facial affective picture system (CFAPS) and found that the average recognition accuracy of the seven emotions is 88.51%, which is 3%～5% higher than the image-based method.

Multi-modal physiological signal emotion recognition based on 3D hierarchical convolution fusion

Wenfen LING, Sihan CHEN, Yong PENG, Wanzeng KONG

2021, 3(1):  76-84.  doi:10.11959/j.issn.2096-6652.202108

Asbtract ( 597 )   HTML ( 58)   PDF (2854KB) ( 927 )   Knowledge map   

Figures and Tables | References | Related Articles | Metrics

In recent years, physiological signals such as electroencephalograhpy (EEG) have gradually become popular objects of emotion recognition research because they can objectively reflect true emotions.However, the single-modal EEG signal has the problem of incomplete emotional information representation, and the multi-modal physiological signal has the problem of insufficient emotional information interaction.Therefore, a 3D hierarchical convolutional fusion model was proposed, which aimed to fully explore multi-modal interaction relationships and more accurately describe emotional information.The method first extracted the primary emotional representation information of EEG , electro-oculogram (EOG) and electromyography (EMG) by depthwise separable convolution network, and then performed 3D convolution fusion operation on the obtained multi-modal primary emotional representation information to realize the pairwise mode local interactions between states and global interactions among all modalities, so as to obtain multi-modal fusion representations containing emotional characteristics of different physiological signals.The results show that the accuracy in the valence and arousal of the two-class and four-class tasks on DEAP dataset are both 98% by the proposed model.

Papers and Reports

Research on human-computer shared control technology based on brain-computer interface

Xiaoyan DENG, Zhuliang YU, Canguang LIN, Zhenghui GU, Yuanqing LI

2021, 3(1):  85-92.  doi:10.11959/j.issn.2096-6652.202109

Asbtract ( 555 )   HTML ( 67)   PDF (1004KB) ( 715 )   Knowledge map   

Figures and Tables | References | Related Articles | Metrics

As an active research direction in the field of neural engineering, brain-computer interface (BCI) has important research significance in biomedicine, neural rehabilitation, intelligent robot and other fields.The human-computer shared control technology based on BCI combines the advantages of human and machine intelligence, and has many advantages compared with the direct control BCI control technology.The human-computer shared control based on non-invasive BCI was classified and introduced from different points of view, and the future research direction and development trend of the human-computer shared control system based on BCI were forecasted to provide new ideas and directions.

Conversion method of magnetic resonance sequence based on detail enhancement

Fan YAN, Yining DI, Jianwei ZHANG, Wei CHEN

2021, 3(1):  93-100.  doi:10.11959/j.issn.2096-6652.202110

Asbtract ( 204 )   HTML ( 17)   PDF (2294KB) ( 543 )   Knowledge map   

Figures and Tables | References | Related Articles | Metrics

Magnetic resonance (MR) is a widely used medical imaging method.However, MR images of certain modalities are not easy to obtain directly and thus need to be converted from other modality.To solve these problem, a conversion method of MR sequence based on detail enhancement was proposed.The proposed method constructed an end-to-end network with two conditional GAN-based modules: resiguals template generation module (ResGAN) and detail enhancement module (EnGAN).The proposed method are tested on a set of registered neurofibroma T1 and STIR images, which demonstrate that the proposed method can restore boundary details and signal intensity details better than other existing methods.

The framework model on internal mechanism of big data intelligent command and control

Zhiqiang HU

2021, 3(1):  101-109.  doi:10.11959/j.issn.2096-6652.202111

Asbtract ( 288 )   HTML ( 37)   PDF (394KB) ( 596 )   Knowledge map   

Figures and Tables | References | Related Articles | Metrics

From the information age to the big data era, the development of information-based command and control to intelligent command and control has become a trend.In view of the challenges faced by the development of intelligent command and control in the era of big data, two new concepts of “intelligent source of soul” and big data “intelligent algorithm viaduct” were proposed.On this basis, the big data intelligent decision-making network was proposed, and then the concept model of big data intelligent command and control was constructed, and the internal intelligent mechanism of intelligent command and control was revealed through the model analysis.Finally, based on the “intelligent simulation engine”, the feasibility of the model research was demonstrated through a simplified model simulation design.

Parallel light field: the framework and processes

Fei-Yue WANG, Xiangbing MENG, Sicong DU, Zheng GENG

2021, 3(1):  110-122.  doi:10.11959/j.issn.2096-6652.202112

Asbtract ( 400 )   HTML ( 32)   PDF (6613KB) ( 637 )   Knowledge map   

Figures and Tables | References | Related Articles | Metrics

The light field is the collection of lights in the environment, and the acquisition, calculation and display of light field information are extremely challenging and complex issues.The ACP-based theory of parallel light field provides a new and effective way to solve this problem.It used the acquired light field information from the actual physical world to construct an enhanced light field of artificial world.And guided by the light field and other information from the actual physical world, light field experiments were conducted in the enhanced light field in all artificial world to obtain the optimal light field acquisition or display schemes.Finally, the parallel optimization of the physical and artificial light fields was established through the process of parallel execution, so that the entire system become a closed-loop system.In this way, intelligent light field acquisition and display were achieved, and a real-virtual theoretical framework for light field information processing and utilization was established.