Please wait a minute...
Most Download
Most Cited
Total visitors:
Visitors of today:
Now online:

Current Issue

    20 December 2019, Volume 1 Issue 4
    Regular Papers
    Variation and learning:fuzzy system and fuzzy inference
    GARIBALDI Jonathan M,Hongyu CHEN,Xiaoshuang LI
    2019, 1(4):  319-326.  doi:10.11959/j.issn.2096-6652.201936
    Asbtract ( 185 )   HTML ( 32)   PDF (1759KB) ( 302 )   Knowledge map   
    Figures and Tables | References | Related Articles | Metrics

    As a decision support system,fuzzy system can deal with uncertainty and has a clear representation of uncertainty knowledge and inference process.But one problem that exists is that computerized decision support systems,including systems that use fuzzy methods,do not have a clear assessment method to determine whether they can be allowed to be used in the real world.A conceptual framework of indistinguishable lines as a key component in evaluating computerized decision support systems was proposed,and some case studies were given.The case proves that the performance of human experts is not perfect,and the fuzzy system can simulate human performance at the technical level,including the variation of human experts.In summary,fuzzy methods are necessary for the representation and reasoning of uncertainty of the knowledge-based systems.Variation is an important form of learning.When evaluating AI systems,imperfect performance should be accepted.

    Development prospect of fuzzy system oriented to interpretable artificial intelligence and big data
    Dewang CHEN,Jijie CAI,Yunhu HUANG
    2019, 1(4):  327-334.  doi:10.11959/j.issn.2096-6652.201937
    Asbtract ( 506 )   HTML ( 63)   PDF (907KB) ( 860 )   Knowledge map   
    Figures and Tables | References | Related Articles | Metrics

    As a universal approximator with strong interpretability,fuzzy system has been widely used in various fields.Although the current theoretical research on fuzzy system is not mature enough,there are still many problems such as too many rules,optimization difficulties,dimension curse,which make it difficult to deal with high-dimensional large data.Although deep neural network has made remarkable progress and can process large data such as image and voice very well,its interpretability is not good and it is difficult to be used in important security-related occasions.Therefore,it is necessary to study an interpretable artificial intelligence algorithm based on fuzzy system.Combining the advantages of deep neural network and fuzzy system,it is possible to solve the problem of high dimensional and large data by studying the deep fuzzy system and its algorithm.The development history and research progress of fuzzy system separately was mainly reviews,and its future development direction according to its existing problems was pointed out,and the summary of this article and the prospect for further research about the problems were given.

    Parallel art:artistic creation under human-machine collaboration
    Chao GUO,Yue LU,Yilun LIN,Fan ZHUO,Fei-Yue WANG
    2019, 1(4):  335-341.  doi:10.11959/j.issn.2096-6652.201938
    Asbtract ( 265 )   HTML ( 34)   PDF (838KB) ( 411 )   Knowledge map   
    Figures and Tables | References | Related Articles | Metrics

    The artistic creation of machine has drawn considerable attention and achieved significant development in recent years.There are more and more artworks processed partially by specific algorithms,or even created entirely by machines.Despite their popularity,it is hard for these artworks to be accepted by humans because of their underwhelming sensory impacts and lack of empathy.However,their impact is highly concerned by the art community.A theoretical framework called the parallel art system was proposed to solve the technical and human-machine-relationship challenges in artistic creation.The system aims to build a partnership between humans and machines,enabling them to collaborate in a parallel manner,and even provide a new way to integrate human emotion and machine logic.

    Application and practice of machine learning model in real-time anti-fraud in the era of digital finance
    Hanping CAO,Xiaojing ZHANG,Ruijie ZHU,Xiaola HUANG
    2019, 1(4):  342-351.  doi:10.11959/j.issn.2096-6652.201939
    Asbtract ( 483 )   HTML ( 48)   PDF (3227KB) ( 455 )   Knowledge map   
    Figures and Tables | References | Related Articles | Metrics

    In recent years,with the rapid development of FinTech,digital finance has flourished and brought huge positive effect on society.Meanwhile,new risks have been introduced into banks.For example,the black production related to network security has experienced explosive growth,and telecommunication network fraud has caused property losses to the public.In the era of digital finance,the commercial banks have not only ushered in new opportunities and dynamics,but also faced new challenges and requirements for digital transformation.As a result,e-finance has become a new battlefield.With this context,a real-time anti-fraud machine learning model based on high-dimensional transaction behavior portrait through enhanced RFM feature-derivation and machine learning modeling was established in this paper.Relying on the new technologies such as big data,stream computing,a model application solution to real-time risk control was formed including systematic deployment,model application strategies and iterative model optimization.Through practical observation,the AUC of the model reaches 0.972,which provides a keen insight into fraud risk,realizes millisecond-level risk identification,and promotes risk control ability of e-finance significantly.

    Digital monitoring and modeling of activities:the IVFC case study
    Xingkai SUN,Xiao WANG,Hao LU
    2019, 1(4):  352-368.  doi:10.11959/j.issn.2096-6652.201940
    Asbtract ( 147 )   HTML ( 10)   PDF (8244KB) ( 161 )   Knowledge map   
    Figures and Tables | References | Related Articles | Metrics

    At present,the frequency of various activities,such as competitions and conferences is increasing day by day.However,there is a lack of an effective platform of digital monitoring and analysis of online media for such events.Based on the method framework in parallel intelligence and knowledge automation theory,a methodological framework for network media monitoring and modeling analysis for events was proposed in this paper,and takes the “Intelligent Vehicle Future Challenge” as an application example.The dataset was collected from three main types of online media (news,WeChat,Weibo) data from 2009 to 2017.Then the online media data were analyzed and visualized from the multiple dimensions including time and space,hotspots,release sources,keywords,topics,semantics,and entities.The results show that the relevant framework can provide effective digital monitoring means and auxiliary decision support for the activity-related organizations.

    Brain network evolution modeling based on Alzheimer’s disease
    Bingjie NI,Wei LI,Xi CHEN
    2019, 1(4):  369-378.  doi:10.11959/j.issn.2096-6652.201941
    Asbtract ( 309 )   HTML ( 40)   PDF (3491KB) ( 244 )   Knowledge map   
    Figures and Tables | References | Related Articles | Metrics

    As a common and frequently-occurring disease,Alzheimer’s disease seriously affects the quality and level of life of the middle-aged and elderly.An in-depth understanding of the pathogenesis and progression of Alzheimer’s disease is important for the development of prevention and treatment options.Most of the existing studies were based on the static property analysis of the brain network before and after the lesion,and dynamic evolutionary mechanism of the lesion process was usually neglected.A dynamic evolution model was proposed based on the level of brain networks of Alzheimer’s disease by analyzing the longitudinal development to simulate plastic changes of the process.Finally,the rationality of the model was verified by evaluating the evolutionary results from multiple perspectives.The study provides a new idea for early diagnosis,functional evaluation and prediction of Alzheimer’s disease.

    The architecture and scheme of the hybrid-augmented intelligence open innovation platform based on the virtual and real systems
    Jun ZHANG,Lingxi LI,Yilun LIN,Tianyun ZHANG,Ke ZHANG,Peidong XU,Keyu RUAN,Dan SHEN
    2019, 1(4):  379-391.  doi:10.11959/j.issn.2096-6652.201942
    Asbtract ( 231 )   HTML ( 8)   PDF (1374KB) ( 372 )   Knowledge map   
    Figures and Tables | References | Related Articles | Metrics

    The overall goal and problems of the construction of the hybrid-augmented intelligent open innovation platform were first put forward,and a virtual and real systems driven platform architecture scheme was proposed.Then the basic form of the platform was described,and the mechanism and function of modules at all levels in the platform were elaborated,the key technologies were also involved,including human-computer interaction technology,data processing technology and digital virtual industry technology.Finally,the guarantee and incentive mechanism of open innovation was designed,and the application of block chain technology in data security of open innovation platform was discussed.

    Transportation scene recognition based on high level feature representation
    Wenhua LIU,Yidong LI,Tao WANG,Jun WU,Yi JIN
    2019, 1(4):  392-399.  doi:10.11959/j.issn.2096-6652.201943
    Asbtract ( 240 )   HTML ( 21)   PDF (5430KB) ( 238 )   Knowledge map   
    Figures and Tables | References | Related Articles | Metrics

    With the development of intelligent transportation,it has become an urgent problem to quickly and accurately recognize complex traffic scene.In recent years,a large number of scene recognition methods have been proposed to improve the effectiveness of traffic scene recognition,however,most of these algorithms cannot extract the semantic characteristics of the concept of vision,leading to the low recognition accuracy in traffic scenes.Therefore,a novel traffic scene recognition algorithm which extracts the high-level semantic and structural information for improving the accuracy was proposed.A system to discover semantically meaningful descriptions of the scene classes to reduce the “semantic gap” between the high level and the low-level feature representation was built.Then,the multi-label network was trained by minimizing loss function (namely,element-wise logistic loss) to obtain the high-level semantic representation of traffic scene images.Finally,experiments on four large-scale scene recognition datasets show that the proposed algorithm considerably outperforms other state-of-the-art methods.

    Transparent farm based on blockchain technology
    Xiujuan WANG,Jing HUA,Mengzhen KANG,Haoyu WANG,Yong YUAN,Fei-Yue WANG
    2019, 1(4):  400-408.  doi:10.11959/j.issn.2096-6652.201944
    Asbtract ( 234 )   HTML ( 41)   PDF (3420KB) ( 372 )   Knowledge map   
    Figures and Tables | References | Related Articles | Metrics

    With the development of new types of agricultural business mode,consumers can send orders remotely to the growers,in order to satisfy their high requirements for the safety of agricultural products.In addition to consumers and growers,stakeholders of agricultural supply chain include farm managers,agricultural input suppliers,agricultural knowledge suppliers and third-party regulators and financial service providers.Therefore,how to build a trustworthy community,to protect the benefit of all parties and achieve sustainable development are challenges that are faced by new farming business mode.For farm managing mode with multiple stakeholders,smart contracts provide a credit guarantee mechanism.An online farm platform based on smart contracts,which constructs a trustworthy community corresponding to the offline one with blockchain technology was proposed in this paper.Differing from traditional traceability systems,the characteristics of the decentralized data management and non-tamper ability were utilized,and the authenticity of data through the mutual verification could be ensured.These methods increase the cost of data fraud,and thus achieve a trustworthy agricultural community.This work aims at ensuring the data validity in Internet+ agriculture,and supporting the orderly sustainable development of emerging agricultural management mode.

    Quantum blockchain:can blockchain integrated with quantum information technology resist quantum supremacy?
    Jun ZHANG,Yong YUAN,Xiao WANG,Fei-Yue WANG
    2019, 1(4):  409-414.  doi:10.11959/j.issn.2096-6652.201945
    Asbtract ( 417 )   HTML ( 30)   PDF (653KB) ( 386 )   Knowledge map   
    Figures and Tables | References | Related Articles | Metrics

    Quantum computers,which substantially exceed traditional computing speed and data processing capacity,are gradually moving from theory toward practice.The tremendous computing power of quantum computers will bring fundamental challenges to current information encryption mechanism.Two key applications of quantum information technology were introduced,followed by comments on how quantum supremacy threatens the current blockchain consensus and encryption mechanisms.Meanwhile,considering loopholes existing in the current blockchain technology,a blockchain system integrating quantum technology was discussed aiming to prevent the threat of quantum supremacy.

    Federated visualization:a new model for privacy-preserving visualization
    Yating WEI,Zhiyong WANG,Shuyue ZHOU,Wei CHEN
    2019, 1(4):  415-420.  doi:10.11959/j.issn.2096-6652.201946
    Asbtract ( 429 )   HTML ( 36)   PDF (2458KB) ( 490 )   Knowledge map   
    Figures and Tables | References | Related Articles | Metrics

    The concept,architecture,methods and applications of federated visualization were introduced.The federated visualization framework is capable of encrypting and training a visual model that reflect the characteristics of the entire data for specific tasks and scenarios.The federated visualization framework is an extension and application of federated learning,which emphasized using mutual benefit and win-win federal cooperation to visually analyze multi-source data under the premise of ensuring data privacy,towards breaking down data barriers in various fields and industries and realizing the sharing of data and knowledge.

    Personalized recommendation algorithm based on user behavior analysis
    Jun JIA,Bin ZHANG,Zhiyuan LI,Wei WEI,Hao WEI
    2019, 1(4):  421-426.  doi:10.11959/j.issn.2096-6652.201947
    Asbtract ( 262 )   HTML ( 48)   PDF (1225KB) ( 273 )   Knowledge map   
    Figures and Tables | References | Related Articles | Metrics

    With the development of business intelligence system and data mining technology,user behavior data has an important impact on enterprise decision-making.For the network e-commerce platform,the results of these data analysis can be used to push items of interest to specific users,which can enhance the user experience and the business value of the platform.A personalized recommendation algorithm based on user behavior analysis was proposed,which transforms user behavior information into user rating matrix,and an improved regularized nonnegative matrix decomposition algorithm was also proposed,which adds bias information to the original regularized nonnegative matrix decomposition.This algorithm can fully mine the user behavior information such as click,purchase,browse,collect,etc.,and actively push the items of interest to the users.The experimental results verify the effectiveness and efficiency of the algorithm.