Data Mining - Concepts and Techniques.

“Unpaired Image-to-Image Translation using Cycle-Consistent Adversarial Networks”の学習報告

Social Network Analysis

It is shown that, particularly for terrestrial cellular telephony, the interference-suppression feature of CDMA (code division multiple access) can result in a many-fold increase in capacity over analog and even over competing digital techniques. A single-cell system, such as a hubbed satellite network, is addressed, and the basic expression for capacity is developed. The corresponding expressions for a multiple-cell system are derived. and the distribution on the number of users supportable per cell is determined. It is concluded that properly augmented and power-controlled multiple-cell CDMA promises a quantum increase in current cellular capacity. >

On the capacity of a cellular CDMA system

Recommender systems have developed in parallel with the web. They were initially based on demographic, content-based and collaborative filtering. Currently, these systems are incorporating social information. In the future, they will use implicit, local and personal information from the Internet of things. This article provides an overview of recommender systems as well as collaborative filtering methods and algorithms; it also explains their evolution, provides an original classification for these systems, identifies areas of future implementation and develops certain areas selected for past, present or future importance.

http://romisatriawahono.net/lecture/rm/survey/information%20retrieval/Bobadilla%20-%20Recommender%20Systems%20-%202013.pdf

Recommender systems survey

Recent voice activity detection (VAD) schemes have aimed at leveraging the decent neural architectures, but few were successful with applying the attention network due to its high reliance on the encoder-decoder framework. This has often let the built systems have a high dependency on the recurrent neural networks, which are costly and sometimes less context-sensitive considering the scale and property of acoustic frames. To cope with this issue with the self-attention mechanism and achieve a simple, powerful, and environment-robust VAD, we first adopt the self-attention architecture in building up the modules for voice detection and boosted prediction. Our model surpasses the previous neural architectures in view of low signal-to-ratio and noisy real-world scenarios, at the same time displaying the robustness regarding the noise types. We make the test labels on movie data publicly available for the fair competition and future progress.

Self-Attentive VAD: Context-Aware Detection of Voice from Noise

The aim of this study is to recommend relevant information to users by organizing user communities on electronic learning environment. In this paper, we propose a weblog-based approach to modeling users during collaborative learning process. Thereby, we formulate user behaviors on blogosphere, e.g., posting articles, linking to neighbors, and interactions between neighbors. These user models are capable of being compared with others to quantify similarities between users. We apply co-occurrence analysis methods. In this study, we deploy BlogGrid platform to support information pushing service to students. Through our experimental results, we found out that average weighting measurement scheme with co-occurrence patterns from responding (e.g., comments and trackback) activities is the most significant patterns for information pushing on collaborative learning.

Collaborative user tracking for community organization on blogosphere: a case study of eLearning@BlogGrid

As the Internet infrastructure has been developed, E-mail is regarded as one of the most important methods for exchanging information because of easy usage and low cost. Meanwhile, exponentially growing unwanted mails in users' mailbox have been raised as a main problem. To solve this problem, researchers have suggested many methodologies that are based on Bayesian classification. The kind of system usually shows high performances of precision and recall. But they have several problems. First, it has a cold start problem, that is, training phase has to be done before execution of the system. The system must be trained about spam and non-spam mail. Second, its cost for filtering spam mail is higher than rule-based system. Third, if E-mail has only few terms those represent its contents, the filtering performance is fallen. In this paper, we have focused on the last issued problem and we suggest spam mail filtering system using Semantic Enrichment. For the experiment, we tested the performance by using the measurements like precision, recall, and F1-measure. As compared with Bayesian classifier, the proposed system obtained 4.1%, 10.5% and 7.64% of improved precision, recall and Fl-measure, respectively.

Spam mail filtering system using Semantic enrichment

With the increasing popularity of online video, efficient captioning and displaying the captioned text (subtitles) have also been issued with the accessibility. However, in most cases, subtitles are shown on a separate display below a screen. As a result, some viewers lose condensed information about the contents of the video. To elevate readability and visibility of viewers, in this paper, we present a framework for displaying synchronized text around a speaker in video. The proposed approach first identifies speakers using face detection technologies and subsequently detects a subtitles region. In addition, we adapt DFXP, which is interoperable timed text format of W3C, to support interchanging with existing legacy system. In order to achieve smooth playback of multimedia presentation, such as SMIL and DFXP, a prototype system, namely MoNaPlayer, has been implemented. Our case studies show that the proposed system is feasible to several multimedia applications.

Automatic Subtitles Localization through Speaker Identification in Multimedia System

Recently junk e-mail has been one of the most serious information overloading problems. This paper proposes multi-agent system to collaboratively filter spams from users' mail stream. This multi-agent system is organized by personal agents automatically extracting features based on users' manual filtering and facilitator managing knowledge extracted by personal agents. Especially, personal agents can analyze junk e-mails for extracting keyphrases and communicate with the others. Due to the domain specific properties of junk e-mail filtering we have formalized the features extracted from e-mail to be highly understandable and efficiently sharable. Thereby, we have defined two types of features in e-mail as apriori feature and keyphrase-based conceptual one. Besides, these features are integrated in the blackboard system of facilitator for collaborative learning. Finally, we show the filtering performance of collaborative learning by comparing with that of personal agent.

Geun-Sik Jo

Papers

Self-Attentive VAD: Context-Aware Detection of Voice from Noise

Collaborative user tracking for community organization on blogosphere: a case study of eLearning@BlogGrid

Spam mail filtering system using Semantic enrichment

Automatic Subtitles Localization through Speaker Identification in Multimedia System

Collaborative junk e-mail filtering based on multi-agent systems