2018 IEEE International Conference on Data Mining Workshops (ICDMW)最新文献

{"title":"Learning via Social Preference: A Coarse-to-Fine Training Strategy for Style Transfer Systems","authors":"Zhuoqi Ma, N. Wang, Yi Hao, Jie Li, Xinbo Gao","doi":"10.1109/ICDMW.2018.00064","DOIUrl":"https://doi.org/10.1109/ICDMW.2018.00064","url":null,"abstract":"Neural style transfer represented a paradigm shift in image artistic rendering and applications in digital entertainment. Generative adversarial networks are considered as a general solution for style transfer problems. Researchers have explored multiple learning algorithms to increase the learning ability of style transfer networks. However, such research has overlooked an important outside motivator: social preference. Classical style transfer networks are trained offline and worked as a stationary model to transfer photos into stylized images, without any interaction with the environment. Based on the ideas from online training, we propose a new coarse-to-fine training strategy for neural style transfer systems to adapt to social preference change. In coarse stage, a primary model is learned via normal training method. In fine stage, the model is updated with online learning approach and sequentially added new data. We show that our approach exhibits improved performance compared to stationary model from visual effect and reflection of social preference. We conclude that the coarse-to-fine training strategy can improve the output of the generative model in social media environment.","PeriodicalId":259600,"journal":{"name":"2018 IEEE International Conference on Data Mining Workshops (ICDMW)","volume":"15 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121630523","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Anomaly-Based Insider Threat Detection Using Deep Autoencoders","authors":"Liu Liu, O. Vel, Chao Chen, Jun Zhang, Yang Xiang","doi":"10.1109/ICDMW.2018.00014","DOIUrl":"https://doi.org/10.1109/ICDMW.2018.00014","url":null,"abstract":"In recent years, the malicious insider threat has become one of the most significant cyber security threats that an organisation can be subject to. Due to an insider's natural ability to evade deployed information security mechanisms such as firewalls and endpoint protections, the detection of an insider threat can be challenging. Moreover, compared to the volume of audit data that an organization collects for the purpose of intrusion/anomaly detection, the digital footprint left by a malicious insider's action can be minuscule. To detect insider threats from large and complex audit data, in this paper, we propose a detection system that implements anomaly detection using an ensemble of deep autoencoders. Each autoencoder in the ensemble is trained using a certain category of audit data, which represents a user's normal behaviour accurately. The reconstruction error obtained between the original and the decoded data is used to measure whether any behaviour is anomalous or not. After the data has been processed by the individually trained autoencoders and the respective reconstruction errors obtained, a joint decision-making mechanism is used to report a user's overall maliciousness score. Numerical experiments are conducted using a benchmark dataset for insider threat detection. Results indicate that the proposed detection system is able to detect all of the malicious insider actions with a reasonable false positive rate.","PeriodicalId":259600,"journal":{"name":"2018 IEEE International Conference on Data Mining Workshops (ICDMW)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121944760","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"MinerLSD: Efficient Local Pattern Mining on Attributed Graphs","authors":"M. Atzmüller, H. Soldano, G. Santini, Dominique Bouthinon","doi":"10.1109/ICDMW.2018.00040","DOIUrl":"https://doi.org/10.1109/ICDMW.2018.00040","url":null,"abstract":"Local pattern mining on attributed graphs is an important and interesting research area combining ideas from network analysis and graph mining. In this paper, we present MinerLSD, a method for efficient local pattern mining on attributed graphs. In order to prevent the typical pattern explosion in pattern mining, we employ closed patterns for focusing pattern exploration. In addition, we exploit efficient techniques for pruning the pattern space: We adapt a local variant of the Modularity metric with optimistic estimates, and include graph abstractions. Our experiments on several standard datasets demonstrate the efficacy of our proposed novel method MinerLSD as an efficient method for local pattern mining on attributed graphs.","PeriodicalId":259600,"journal":{"name":"2018 IEEE International Conference on Data Mining Workshops (ICDMW)","volume":"11 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121959581","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A Scalable and Extensible Blockchain Architecture","authors":"Yue Yu, Ran Liang, Jiqiu Xu","doi":"10.1109/ICDMW.2018.00030","DOIUrl":"https://doi.org/10.1109/ICDMW.2018.00030","url":null,"abstract":"It has been a long time since Nakamoto Satoshi introduced Bitcoin [1]. However, the underlying technology, blockchain, has been getting more and more attention. It is regarded as one of the most important technology for the fourth industrial revolution, although there exit few the real-world production-level applications due to the hardly tolerant performance of most of the existing blockchain systems. Therefore, how to improve its performance has been identified as one of the most significant research directions in order to use blockchain for practical applications. In this paper, we present a method to build a scalable and extendible blockchain system by multiple chains and sharding techniques.","PeriodicalId":259600,"journal":{"name":"2018 IEEE International Conference on Data Mining Workshops (ICDMW)","volume":"33 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116821046","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"JobSense: A Data-Driven Career Knowledge Exploration Framework and System","authors":"Xavier Jayaraj Siddarth Ashok, Ee-Peng Lim, Philips Kokoh Prasetyo","doi":"10.1109/ICDMW.2018.00200","DOIUrl":"https://doi.org/10.1109/ICDMW.2018.00200","url":null,"abstract":"Today's job market sees rapid changes due to technology and business model disruptions. To fully tap on one's potential in career development, one has to acquire job and skill knowledge through working on different jobs. Another approach is to seek consultation with career coaches who are trained to offer career advice in various industry sectors. The above two approaches, nevertheless, suffer from several shortcomings. The on-the-job career development approach is highly inefficient for today's fast changing job market. The latter career coach assisted approach could help to speed up knowledge acquisition but it relies on expertise of career coaches but experienced career coaches are scarce, and they too require update of jobs and skills knowledge. Meanwhile, with wide adoption of Online Professional Net-works (OPNs) such as LinkedIn, Xing and others, publicly shared user profiles have become a treasure trove of job and skill related data. Job and skill related information is also hidden in the sea of online job posts and ads. Manually exploring and acquiring knowledge from these varieties of information are daunting and time-consuming. On the other hand, one needs substantial effort to personalize the acquired knowledge to his/her career interests. There is a dire need for a self-help tool to ease this knowledge acquisition and exploration problems. Before that, there is also a need to create and maintain a large knowledge base of these jobs, skills and careers. Our data-driven, automated knowledge acquisition and interactive exploration system, JobSense, would help users meet the above challenges. JobSense enables users at several stages of career, to explore this knowledge at ease via interactive search, easy navigation, bookmarking of information entities and personalized suggestions. Also we have introduced a career path generation module, to return relevant career paths to the users.","PeriodicalId":259600,"journal":{"name":"2018 IEEE International Conference on Data Mining Workshops (ICDMW)","volume":"87 3-4 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123432850","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"DGCC: A Case for Integration of Brain Cognition and Intelligence Computation","authors":"Guoyin Wang","doi":"10.1109/ICDMW.2018.00076","DOIUrl":"https://doi.org/10.1109/ICDMW.2018.00076","url":null,"abstract":"Both human brain and computer (electronic brain) could process data and do some cognition and computation tasks. Is the cognition of human brain equal to the computation of computer? It is obviously not. In this talk, the relationships of brain cognition models and intelligence computation models are summarized into four different types, that is, human brain cognition inspired intelligence computation (BCIIC), intelligence computation without human brain cognition (ICOBC), intelligence computation assisted human brain cognition (ICABC), and the integration of human brain cognition and intelligence computation (BC&IC). There are three paradigms in traditional artificial intelligence (AI) studies, that is, symbolism AI, connectionism AI, and behaviorism AI. The physical symbol system hypothesis is used in the symbolism AI. Human brain cognition is taken as a kind of symbolic processing, and the processes of human thinking are computed by symbol in the symbolism AI [1,2]. The connectionism AI relies on the bionics to simulate human brain. In the connectionism AI, neuron is taken as the basic unite of human thinking, and the intelligence is taken as the result of interconnected neurons competition and collaboration [3,4]. In the behaviorism AI, intelligence depends on the perception and behavior, “Perception-action” model is used, and intelligence may not require knowledge, knowledge representation and knowledge reasoning [5]. The symbolism AI and connectionism AI are two different types of human brain cognition inspired intelligence computation, while the behaviorism AI is a representative case of intelligence computation without human brain cognition. Usually, AI researchers get some inspiration from human brain cognition in their studies. On the other way, intelligence computation could also assist human brain cognition studies. The bidirectional cognitive computing model (BCC) is such a case. It studies the bidirectional transformations between the intension and extension of a concept. It is used to simulate some human brain cognition tasks such as learning and recognition [6,7]. Cognitive computing is one of the core fields of artificial intelligence [8,9]. Data-driven granular cognitive computing (DGCC) is an example of the integration of human brain cognition and intelligence computation [10,11]. It takes data as a special kind of knowledge expressed in the lowest granularity level of a multiple granularity space. It integrates two contradictory mechanisms, namely, the human’s cognition mechanism of ‘‘global precedence’’ which is a cognition process of ‘‘from coarser to finer’’ and the information processing mechanism of machine learning systems which is ‘‘from finer to coarser’’, in a multiple granularity space. It is data-driven cognitive computing model. The integration of human brain cognition and intelligence computation would be an important research topic of artificial intelligence. Some scientific research issues of the integration of ","PeriodicalId":259600,"journal":{"name":"2018 IEEE International Conference on Data Mining Workshops (ICDMW)","volume":"109 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123738385","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A Novel Learning Approach to Improve Mobile Application Recommendation Diversity","authors":"Kun Niu, Haizhen Jiao, Xiao Xu, Cheng Cheng, Chao Wang","doi":"10.1109/ICDMW.2018.00185","DOIUrl":"https://doi.org/10.1109/ICDMW.2018.00185","url":null,"abstract":"With the popularity of smart phones, plenty of mobile phone applications are developed to meet people's various needs, and mobile application recommendation has become a popular and challenging topic. Most studies focus on learning user preferences from various information both on user-side and APP-side, and recommending based on user similarity or app similarity. However, these methods all have a high probability to cause serious homogenization problems that can not meet users' unknown/new needs. Therefore, recommending diverse apps is more likely to cover users' all preferences, and even guide users to discover new needs and interests. To this end, we give the definition of Application Diversity that taking into account the similarity between apps and the relevance of categories, and propose a novel application recommendation approach that consists of two parts, P-Stair Neural Network (P-SNN) and Dynamic Adjustment Method (DAM). First, P-SNN learns user preferences from multi-dimensional data by using deep neural networks techniques, and predicts users' ratings for uninstalled applications. Then, DAM selects TOP-N applications as the final recommend list with considering both user preferences and recommend diversity. Several experiments on different datasets shows that our algorithm effectively improves the diversity of recommendations in the case of similar accuracy.","PeriodicalId":259600,"journal":{"name":"2018 IEEE International Conference on Data Mining Workshops (ICDMW)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129583237","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A Graph-Based Approach for Learner-Tailored Teaching of Korean Grammar Constructions","authors":"Mikhail Akimov, Ekaterina Loginova, Maxim Musin","doi":"10.1109/ICDMW.2018.00057","DOIUrl":"https://doi.org/10.1109/ICDMW.2018.00057","url":null,"abstract":"Foreign language learning on an intermediate level is often a complicated task, as it requires acquisition not only of vocabulary and language rules but of context-dependent meanings of words. This is especially relevant for Category IV languages like Korean, in which the same tokens could be both words and grammar tags. The textbook adapted versions of words and contexts often fail to capture the existing complexity, while the real world examples may be too hard for a novice and even an intermediate level learner. In addition, the particular learner may be familiar with some functions and contexts for a particular word, but not with the other ones. To alleviate this complexity problem, we propose a semantic graph based personalized tutoring system. The learning corpus is constructed using real-world sentences from a newspaper, which are translated using an automated service and processed with NLP techniques to extract token functions. A graph is used to track word and grammar construct context and thus find similar and dissimilar word use cases, as well as for the estimation of sentence complexity. The system then shows words and grammar constructs from real-world sentences to learners and records their understanding in each context. The collected context dependent understanding data together with the sentence complexity estimation are then used to estimate the learner's level and tailor the sentence set accordingly. The resultant approach could be extended to the tutoring of context-dependent meanings in other languages.","PeriodicalId":259600,"journal":{"name":"2018 IEEE International Conference on Data Mining Workshops (ICDMW)","volume":"3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125284052","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A Domain-Independent Text Segmentation Method for Educational Course Content","authors":"Yuwei Tu, Ying Xiong, Weiyu Chen, Christopher G. Brinton","doi":"10.1109/ICDMW.2018.00053","DOIUrl":"https://doi.org/10.1109/ICDMW.2018.00053","url":null,"abstract":"In this study, we have proposed a domain-independent text segmentation algorithm which is particularly useful in online educational courses. Text segmentation is proven to be helpful in improving the readability of large corpora of documents, which is essential in education scenarios. While existing domain-dependent text segmentation methods have much better performance than domain-independent methods in most cases, only domain-independent methods are applicable to sparse training content in education scenarios. Our method, unlike other domain-dependent text segmentation methods, doesn't require heavy training on prior documents, but only need to train on the current corpus of documents with topic distributions and word vector representations. Our proposed method develops text boundaries between small text units in three steps. We first calculate input text features via topical distributions (latent Dirichlet allocation) and word embeddings (GloVe). We then calculate similarity values between such textual features and detect distribution changes between the similarities. We finally perform clustering on the similarities and detect sub-topic boundaries via cluster differences. We test our method on two datasets, one from an online education course and one from a popular public dataset - Choi Dataset. The results demonstrate that our method outperforms other state-of-the-art domain-independent text segmentation approaches while achieving performance comparable to a few domain-dependent algorithms.","PeriodicalId":259600,"journal":{"name":"2018 IEEE International Conference on Data Mining Workshops (ICDMW)","volume":"52 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125523060","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Sequential Variational Learning of Dynamic Factor Mixtures","authors":"A. Samé","doi":"10.1109/ICDMW.2018.00105","DOIUrl":"https://doi.org/10.1109/ICDMW.2018.00105","url":null,"abstract":"The clustering of panel data remains a challenging problem, considering their dynamic and potentially massive nature. The massive aspect of panel data can be related to their number of observations and/or their high dimensionality. In this article, a new model and its estimation method are initiated to tackle these problems. The proposed model is a mixture distribution whose components are dynamic factor analyzers. The model inference, which cannot be performed exactly by classical methods, is realized in the sequential variational framework. In particular, it is established that the proposed algorithm converges in the sense of stochastic gradient algorithms toward an average lower variational bound. Experiments conducted on simulated data illustrate the good practical behavior of the method.","PeriodicalId":259600,"journal":{"name":"2018 IEEE International Conference on Data Mining Workshops (ICDMW)","volume":"19 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128809194","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}