Research by faculties and students in the Computer Industry Laboratory has been carried out in the fields of both basic computer science and industrial applications, such as Artificial Intelligence, Semantic Web Service, Web Data Mining, e-Business, Signal processing, Simulation engineering, and Functional Safety, and environmental impact analysis of energy industry.
Several articles were published as book chapters, journal-contributions, proceedingscontributions in conferences, and technical reports.
The research activities of the Computer Industry Laboratory include the following topics:
Signal processing for plant monitoring and control
Dynamical system identification and analysis
Inverse problem solving for Non-destructive evaluation
Traffic simulation for emergency response management
Functional safety of embedded system [Incheon PAIK] ffl Web Service
Workflow and Business Process Management
Web Data Mining
Semantic e-Business Application
Environmental impact of energy industry
Environmental economics and computer simulation
Assessment of new technologies in electric power generation and consumption
Ryoichi Urayama, Tetsuya Uchimoto, Toshiyuki Takagi, and Shigeru Kanemoto. Quantitative Evaluation of Pipe Wall Thinning by Electromagnetic Acoustic Resonance. E-journal of Advanced Maintenance (EJAM), http://www.jsm.or.jp/ejam/, 2010.
In this study, Electromagnetic Acoustic Resonance method (EMAR) is applied to online monitoring of pipe wall thinning, and accuracy and stability of measurement is discussed through field tests using a large-scale corrosion test loop. Although EMAR has a high capability of evaluating thickness or ultrasonic velocity, targets are usually limited to ones with simple geometry such as plates. In order to measure thickness of pipes with complicated wall thinning, superposition of nth compression for data processing method is applied to extract thickness information from spectral responses. Results of two-month monitoring show that EMAR with the proposed signal processing method has a capability of evaluating pipe wall thinning with the accuracy of several ?m at high temperature of 165 degrees C.
Ihor Lubashevsky and Shigeru Kanemoto. Scale-free memory model for multiagent reinforcement learning: Mean field approximation and rockpaper-scissors dynamics. NEuropean Physical Journal ser. B, 2010.
A continuous time model fro multiagent systems governed by reinforcement learning with scale free memory is developed. The agents are assumed to act independently of one another in optimizing their choice of passible actions via trial-and-error search. To gain awareness about the action value the agents accumulate in their memory the rewards obtained from taking a apecific action at each value at each moment of time.
S. Kanemoto. Acoustic monitoring of rotating machine by advanced signal processing technology. Nuclear Safety and Simulation, 1(1):22-29, 2010.
The acoustic data remotely measured by handy type microphones are investigated for monitoring and diagnosing the rotational machine integrity in nuclear power plants. The plant operator's patrol monitoring is one of important activities for condition monitoring. However, remotely measured sound has some difficulties for precise diagnosis or quantitative judgment of rotating machine anomaly, since the measurement sensitivity is different in each measurement, and also, the sensitivity deteriorates comparing with an attached type sensor. Hence, in the present study, several advanced signal processing methods are examined and compared in order to find optimum anomaly monitoring technology from the viewpoints of both sensitivity and robustness of performance. The dimension of pre-processed signal feature patterns are reduced into two-dimensional space for the visualization by using the standard principal component analysis (PCA) or the kernel based PCA. Then, the normal state is classified by using probabilistic neural network (PNN) or support vector data description (SVDD). By using the mockup test facility of rotating machine, it is shown that the appropriate combination of the above algorithms gives sensitive and robust anomaly monitoring performance.
Incheon Paik and Hiroshi Mizugai. Recommendation System UsingWeighted TF-IDF and Naive Bayes Classifiers on RSS Contents. Journal of Advanced Computational Intelligence and Intelligent Informatics, 14(6):631 637, 2010.
A recent increase in RDF Site Summary (RSS) feeds, used for news updates and blogs, has been caused by the widespread use of blogs. This means that much effort is now needed to search the contents of RSS feeds because of this enormous quantity of material. To solve this problem, recommendation systems enable users to obtain relevant RSS contents easily and quickly. In previous research, an RSS recommendation system was proposed that used the similarity between the Term Frequency (TF) of the RSS contents and the TF derived from the contents of the user's browsing history for RSS feeds. In this paper, we use Term Frequency-Inverse Document Frequency (TF-IDF) calculations to propose a Weighted TF-IDF method, which focuses on the terms folded by the title tags in RSS contents as characteristic terms. In addition, we propose a new recommendation method, which uses a Naive Bayes classifier in a Machine Learning-based approach. Via experiments, we compare the proposed methods and the existing method in a prototype recommendation system, and we show that the proposed methods outperform the existing method with respect to several evaluation measurements.
Aiguo He, Shigeru Kanamoto, Takenobu Kazuma, Takaaki Furukawa, Shiyang Wang, and Yo ichi Kawai. Operator Tracking System using Particle Filter for Operation Skill Evaluation in Plant Control Room. In The 3rd International Symposium on Symbiotic Nuclear Power Systems for the 21st Century (ISSNP2010), Harnine, China, August 2010.
This paper proposes an automated operator tracking system using particle filter and image processing for helping operator skill evaluation in nuclear power plant operator training facilities. In each control room of such a training facility, a full-scope plant simulator with mock-up control panels is used for real time operator training based on various kinds of accident scenarios. Here, operator behaviors and plant events are recorded, as video-log and event-log, by a multi-media recording system, and utilized for evaluating operator's skills at training review meetings where the instructors and the trainees discuss each trainee's action in the training process. Multiple cameras mounted on the control room ceiling are used for the operator behavior recording. However, the view from these cameras is not a so comfortable interface in order to evaluate how each operator approaches the target panel in appropriate timing corresponding to the plant event. For example, the instructors have to estimate operator's real world position from the views; and in some case other operators might hide the target operator from the cameras. The purpose of the automated operator tracking system is to help the instructors and the operators to evaluate whether the operator position and timing is appropriate in each event or alarm occurrence, by tracking the operators from the recorded video sequence. To attain this, we introduced art image processing technology. Particle filter is one of convenient algorithm for operator tracking. In this algorithm, the main issue is how to recognize the multiple target operators from the background image and get their positions in the coordinate of the control room. For this purpose, one 3D particle filter is used for each operator wearing a colored vest and the similarity calculation algorithm is based on the color histogram. The particles are directly placed inside the control room. By converting particle coordinate into camera coordinate and using the distance between the particle and the camera, the position and size of candidate regions on the video image for similarity calculation can be easily calculated. Camera parameters used in the coordinate conversion are identified by using mesh type markers set on the floor. The above approach is verified by using actual training environment in the BWR operator training center corporation facility.
Shigeru Kanamoto, Aiguo He, Takenobu Kazuma, Shiyang Wang, and Yo ichi Kawai. OPERATOR TRACKING SYSTEM USING PARTICLE FILTER FOR OPERATION SKILL EVALUATION IN PLANT CONTROL ROOM. In Seventh American Nuclear Society International Topical Meeting on Nuclear Plant Instrumentation, Control and Human-Machine Interface Technologies(NPIC&HMIT 2010), Las Vegas, Nevada, November 2010.
This paper proposes an automated operator tracking system using particle filter and image processing for helping operator skill evaluation in nuclear power plant operator training facilities. Here, a full-scope plant simulator with mock-up control panels is used for real time operator training based on various kinds of emergency accident scenarios, and, the operator behaviors are evaluated by instructors. The operator behaviors and plant events are recorded by a multi-media recording system as videoand event-log, and utilized for evaluating the validity of operator's actions in the review meetings between the instructors and trainees. Although multiple camera images are recorded for this purpose, the view from these cameras is not a so comfortable interface to evaluate how each operator approaches the target panel in appropriate timing corresponding to the plant event. So, the present paper tries to provide more comfortable interface by using up-to-date image processing technologies to help the instructors and the operators to check whether the operator position and timing is appropriate in each event or alarm occurrence. To attain this, we introduced the particle filter and the human image recognition technologies. The particle filter is used to model the operator dynamic behavior in 3D real world and to estimate his real position from noisy observation data. The automated human image recognition is also important technology to discriminate the target human image from the background images. The likelihood of target human image is evaluated by using image feature extraction and similarity calculation, and used in the particle filter re-sampling algorithm. Camera parameters used in the coordinate conversion between 2D images and 3D real world are identified by using mesh type markers set on the floor. The performance of the proposed automated tracking algorithms is verified by using an actual operator training environment in the BWR operator training center corporation.
Incheon Paik, Haruhiko Takada, and Wuhui Chen. Transformation of Semi-Structured Abstract Non-Functional Properties for Service Composition. In NFPSLAM-SOC Committee, editor, Proceedings of The 3rd Workshop on NFPSLAM-SOC'10, page Online Version, Naya, Cyprus, November 2010. NFPSLAM-SOC Organizing Committee, LNCS.
Functional properties are discovered at the discovery stage and nonfunctional properties (NFP) are used mainly in the selection stage for service composition. Abstractness of NFPs has to be solved to be used in the selection stage for seamless composition. In our initial work, three levels of abstractness of NFPs (abstract, intermediate, and concrete) and transformation from intermediate level to concrete level were suggested. !!This paper complements the translation from abstract level to intermediate level toward complete transformation of NFPs. To solve vagueness of abstract NFPs, we adapt approaches based on not only ontology but also term similarity. Structure of NFPs is based on the definition of NFPs abstractness level, which has non-terminal compound terms. To evaluate effectiveness of term similarity metrics, vector-based and large corpus based approaches are investigated. Transformation performance based on precision over our test data set and ontology is evaluated.
Incheon Paik, Shisuke Mori, and Wuhui Chen. Semantic Words Similarity in Triple Relation Using Intermediate Concept by PLSI. In IEEE SMC Orgainizing Committee, editor, Proceedings of the IEEE International Conference on Systems, Man, and Cybernetics, Istanbul, Turkey, October 2010. IEEE SMC Orgainizing Committee, IEEE Publishing Service.
Semantic similarity measures play important roles in information retrieval and natural language processing. Several researches calculate semantic similarity between two words using web search engines as corpus or manually compiled corpus. In this paper, a method to find the word Ri between two words P and Q and extract a relation of the words with PLSI (Probabilistic Latent Semantic Indexing) is proposed. The results of the experiments show that using the PLSI with smaller latent class such is effective in getting Ri which is more related to P and Q, and using the PLSI with over 5 latent class is effective in getting veiled relation between P and Q.
Incheon Paik and Wuhui Chen. Design of User Interface for Automatic Service Composition. In IEEE SMC Orgainizing Committee, editor, Proceedings of the IEEE International Conference on Systems, Man, and Cybernetics, Istanbul, Turkey, October 2010. IEEE SMC Orgainizing Committee, IEEE Publishing Service.
Automatic Service Composition (ASC) supports creation of a new value-added composite service from the existing services with automatic manner. There are two approaches, machine-oriented and human-oriented, for the composer. In the machineoriented composer, every composition step is managed by the composer mainly. However, there are many possibilities of interventions by users for better composition performance. In this paper, design of user interface for the machine-oriented ASC based on our new composition architecture is suggested. Possible interactions at the all the stages of service composition are analyzed on the new architecture. Ontologies for ASC UI, visual component, data, and workflow are designed. The SelectorUI is demonstrated as an example of UI for ASC, and whole composition scenario is illustrated. The design paradigm of ASC UI presented in the paper can be applied to human-oriented composition too.
Incheon Paik, Eigo Fujikawa, and Sangkyung Kim. Aggregating Web Service Matchmaking Variants Using Web Search Engine and Machine Learning. In ISAC 2010 Committee, editor, Proceedings of ISAC 2010, Tainan,
Taiwan, Novebmer 2010. ISAC 2010 Organizing Committee, IEEE Publishing Service. Variety of Web Service discovery algorithms had been investigated for improvement of the retrieval quality. Combining the several algorithms according to their strong points, is proposed as enabling more refined discovery consequence. Now, many researches as OWL-Mx are sited as examples, had already shown the method that join together and conclude for the specific domain. However, there are no way to conclude multi-algorithms results. Klusch  shows the brand-new way that leads the conclusion by using machine-learning algorithm Support Vector Machine (SVM)
. In this research, we attempted to apply the SVM aggregation and several new discovery algorithm using similarity based on search engine, shown on Trip Domain service discovery. And, 88 percent over score of precision, were gotten as the result from specifically prepared queries for Trip Domain. This experiment also had shown 10 percent missing which occurred by using web page count based similarity computation. In future work, we will conduct some comparison for getting more reliability of this proposed method.
Hiroshi Mizugai, Incheon Paik, and Wuhui Chen. Scalable Orchestration Strategy for Automatic Service Composition. In IEEE CIT Orgainizing Committee, editor, Proceedings of the IEEE International Conference on Computer Information Technology (CIT'2010), pages 1474-1479, Bradford, UK, June 2010. IEEE CIT Orgainizing Committee, IEEE Publishing Service.
The goal of Automatic Service Composition (ASC) is to create value added services automatically from the existing services. Most researches for service composition are based on four stages (planning, discovery, selection, and execution) and its variation, and they don't consider nested dynamic services that ASC calls the inner ASCs internally in the composition time. In practice, there are many situations which dynamic services and existing services are combined each other in ASC. This research proposes a strategy for orchestrating not only the existing four-staged composition but also nested dynamic service composition. Two types of approach are considered: Top-down and Button-up. Top-down composition can create new services with management by composer as a main, while Bottom-up composition creates a new service with human oriented direction. For the two proposed solutions, the core procedures were described, and the prototype systems were implemented in one domain. Finally, the performance of Top-down composition was evaluated.
K. Ofuji. Improvement of Kalman-Filter Tank Model for Hydropower Output Estimation. In Institute of Statistical Mathematics, Inverse Problem Workshop, Tokyo, November 2010. Institute of Statistical Mathematics(ISM), ISM.
K. Ofuji, April 2009-2010.
Member, Asset Management Survey committee
Incheon Paik, Apr. 2010.
Technical Program Committee Member, nternational Workshop on Benchmarking of XML and Semantic Web Applications (Xbench 2010)
Incheon Paik, June 2010.
Steering Committee Member, EEE International Conference on Computer and Information Technology (CIT10)
Incheon Paik, May 2010.
Editorial Borad, Journal of Information Processing Systems of Korea Information Processing Society
Incheon Paik, Apr. 2010.
Reviewer, IEEE Transactions on Service Computing
Takahito Makabe. Graduation Thesis: Analysis of Forward Contracts of Japan Electric Power Exchange, University of Aizu, 2010.
Thesis Adviser: K. Ofuji
Ryusuke Tashiro. Graduation Thesis: Automatic Service Composition System with Non-Functional Property Transformation, University of Aizu, 2010.
Thesis Adviser: I. Paik
Kyubo Kim. Master Thesis: Information Infrastructure for Innovative Product Design Using Semantic TRIZ, Graduate School, University of Aizu, 2010.
Thesis Adviser: I. Paik
Tetsuya Tashiro. Graduation Thesis: Workflow Generation by Planner for Service Composition Using Ontology, University of Aizu, 2010.
Thesis Adviser: I. Paik
Shohei Sugai. Graduation Thesis: Extracting Related Concept Using PLSI and Its Evaluation by Words Similarity, University of Aizu, 2010.
Thesis Adviser: I. Paik
Eri Koizumi. Graduation Thesis: Ontological Identification of Patterns for Choreographing Business Workflow, University of Aizu, 2010.
Thesis Adviser: I. Paik