Curriculum Vitaes

Kiyoki Yasushi

  (清木 康)

Profile Information

Affiliation
Professor (Distinguished professor, Keio University), Data Science, Musashino University
Degree
工学(慶應義塾)
工学(Keio University)

Researcher number
10169956
J-GLOBAL ID
200901089282037598
researchmap Member ID
5000069496

External link

Papers

 165
  • Xing Chen, Yasushi Kiyoki
    Frontiers in Artificial Intelligence and Applications, Jan 16, 2024  
    This paper aims to analyze the phenomenon of inapplicability of experience, which means that sometimes we make mistakes when we use our past experience to solve current problems. We propose a knowledge model based on the concept of “dark-matter”, which is a term used to describe the time-related data that is hidden from our observation. We use a two-dimensional matrix to represent both time-related and non-time-related data, and we call it space. We also introduce the concept of parallel spaces, which are composed of several spaces that can explain different situations and outcomes. We use case studies to illustrate how knowledge is generated and expressed using “dark-matter” and parallel spaces. We also reveal the reason for the inapplicability of experience and suggest some solutions. The contribution of this paper is that we provide a new perspective and a new model to understand and process knowledge based on “dark-matter” and parallel spaces.
  • Yasushi Kiyoki, Asako Uraki, Shiori Sasaki, Yukio Chen
    Frontiers in Artificial Intelligence and Applications, Jan 16, 2024  
    “Semantic space creation” and “distance-computing” are basic functions to realize semantic computing for environmental phenomena memorization, retrieval, analysis, integration and visualization. We have introduced “SPA-based (Sensing, Processing and Actuation) Multi-dimensional Semantic Computing Method” for realizing a global environmental system, “5-Dimensional World Map System”. This method is important to design new environmental systems with Cyber-Physical Space-integration to detect environmental phenomena occurring in a physical-space (real space). This method maps those phenomena to a multi-dimensional semantic-space, performs semantic computing, and actuates the semantic-computing results to the physical space with visualizations for expressing environmental phenomena, causalities and influences. As an actual system of this method, currently, the 5D World Map System is globally utilized as a Global Environmental Semantic Computing System, in SDG14, United-Nations-ESCAP: (https://sdghelpdesk.unescap.org/toolboxes). This paper presents a semantic computing method, focusing on “Time-series-Analytical Semantic-Space Creation and Semantic Distance Computing on 5D World Map System” for realizing global environmental analysis in time-series. This paper also presents the time-series analysis of actual environmental changes on 5D World Map System. The first analysis is on the depth of earthquakes Earthquake with time-series semantic computing on 5D World Map System, which occurred around the world during the period from Aug. 23rd to Aug. 28th, 2014, and Jan 7th to Jan. 13th, 2023. The second is the experimental analysis of the time-series difference extraction on glacier melting phenomena in Mont Blanc, Alps, during the period from 2013 to 2022, and Puncak Jaya (Jayawijaya Mountains), Papua, during the period from 1991 to 2020 as important environmental changes.
  • Yasuhiro Hayashi, Yasushi Kiyoki, Yoshinori Harada, Kazuko Makino, Seigo Kaneoya
    Frontiers in Artificial Intelligence and Applications, Jan 16, 2024  
    This paper proposes a spatio-temporal and categorical correlation computing method for induction and deduction analysis. This method is a data analytics method to reveal spatial, temporal, and categorical relationships between two heterogeneous sets in past events by correlation calculation, thereby finding insights to build new connections between the sets in the future. The most significant feature of this method is that it allows inductive and deductive data analysis by applying context vectors to compute the relationship between the sets whose elements are time, space, and category. Inductive analysis corresponds to data mining, which composes a context vector as a hypothesis to extract meaningful relationships from trends and patterns of past events. Deductive analysis searches past events similar to a context vector’s temporal, spatial, and categorical conditions. Spatio-temporal information about the events and information such as frequency, scale, and category are used as parameters for correlation computing. In this method, a multi-dimensional vector space that consists of time, space, and category dimensions is dynamically created, and the data of each set expressed as vectors is mapped onto the space. The similarity degree of the computing shows the strength of relationships between the two sets. This context vector is also mapped onto the space and is calculated distances between the context vector and other vectors of the sets. This paper shows the details of this method and implementation method and assumed applications in commerce activities.
  • Yasushi Kiyoki, Koji Murakami, Shiori Sasaki, Asako Uraki
    Frontiers in Artificial Intelligence and Applications, Jan 14, 2022  
    Semantic space creation and computing are essentially significant to realize semantic interpretations of situations and symptoms in human-health. We have presented a semantic space creation and computing method for domain-specific research areas. This method realizes semantic space creation with domain-oriented knowledge and databases. This paper presents a semantic space creation and computing method for “Human-Health Database” with the implementation process for “Human-Health-Analytical Semantic Computing”. This paper also presents a new knowledge base creation method for personal health data for preventive care and potential risk inspection with global and geographical mapping and visualization in 5-Dimensional World Map System. This method focuses on the analysis of personal health and potential-risk inspection and realizes a set of semantic computing functions for semantic interpretations of situations and symptoms in human-health. This method is applied to “Human-Health-Analytical Semantic Computing” to realize world-wide evaluation for (1) multi-parameterized personal health data, such as various biomarkers, clinical physical parameters, lifestyle parameters, other clinical/physiological or human health factors, etc., for health monitoring, and (2) time-series multi-parameterized health data in the national/regional level for global analysis of potential cause of disease. This Human-Health-Analytical Semantic Computing method realizes a new multidimensional data analysis and knowledge sharing for a global-level health monitoring and disease analysis. The computational results are able to be visualized in the time-series difference of the values in each place, the difference between the values of multiple places in a focused area, and the time-series differences between the values of multiple places to detect and predict a potential-risk of diseases.
  • Koichiro Kawashima, Yasuhiro Hayashi, Yasushi Kiyoki, Tetsuya Mita
    Frontiers in Artificial Intelligence and Applications, Jan 14, 2022  
    Dynamic routing with combinations of mobility and activity is expected as new methodology for supporting sensitivity to various contexts for traveling. It is important to realize the dynamics by integrating “mobility and activity” in physical and cyber spaces. This paper presents a mobility and activity integration system for making routing plans from an original point to a destination with a scenario as “sensitivity to context” on the route. The “sensitivity to context” expresses reactions to the intentions and situations of a moving user. This system applies semantic computing to find out the appropriate mobility and activity, that dynamically calculates semantic associations between user’s intentions and mobility services. This system makes a moving plan reflecting “sensitivity to context” created by query creation operators for synthesizing and expressing “everyday intention” and “mobility situation.” This system has a distance calculation function for “feature value vectors” expressing the means of mobility and the features of facility spots, and outputs some expected means of moving towards the destination with activities on the route.

Misc.

 34
  • 川島宏一郎, 川島宏一郎, 林康弘, 清木康, 三田哲也
    日本データベース学会和文論文誌(Web), 20-J, 2022  
  • Virach Sornlertlamvanich, Petchporn Chawakitchareon, Aran Hansuebsai, Chawan Koopipat, Bernhard Thalheim, Yasushi Kiyoki, Hannu Jaakkola, Naofumi Yoshida
    Frontiers in Artificial Intelligence and Applications, 301 v-vi, 2018  
  • Yoshiko Itabashi, Shiori Sasaki, Yasushi Kiyoki
    じんもんこん2013論文集, 2013(4) 71-78, Dec 5, 2013  
  • Nguyen, D. T. N, Sasaki, S, Kiyoki, Y
    International Forum for Sustainable Asia and the Pacific (IGES - ISAP 2012), Jul, 2012  Peer-reviewed
  • NGUYEN Diep Thi-Ngoc, SASAKI Shiori, KIYOKI Yasushi
    教育システム情報学会研究報告, 27(2) 9-12, Jul, 2012  
  • 佐々木 史織, 小杉 尚子, 清木 康
    研究報告データベースシステム(DBS), 2011(27) 1-8, Oct, 2011  
    フィジカル・エクササイズにおいて,トレーナーが運動目的に合致する楽曲を選択する,または運動をする人が個人特性 (身体・嗜好) に合った楽曲を選択することの支援を目的として,実施する運動の特性と個人特性に合った楽曲を検索し,運動の負荷に合わせて楽曲の再生を調整するシステムの実現方式について述べる.本システムの特徴は,運動の特徴を記述する "運動属性" と,運動に強く関連する楽曲の特徴量 (リズム,テンポ等) を表す "楽曲属性" の関連性を定義する運動・楽曲特徴変換マトリクスを設定し,運動と個人特性に応じて楽曲データベース中から適切な楽曲群を選択する点にある.This paper describes architecture and a function set of a music recommendation system for physical exercise with users' characteristics adaptation functions. The system retrieves music tunes which fit both exercises and users' preferences and adjusts the tunes based on the exercise load. The system will help trainers and people who do exercise so that they can find tunes for their exercises. The features of the system are 1) making an exercise-music transformation matrix that defines relationship between the exercise attributes based on the exercise features and music attributes that strongly relate to exercises, for example rhythm and tempo, and 2) retrieving tunes based on the exercises and users' preferences from a music database.
  • Sasaki, S, Kiyoki, Y
    National Electronic and Computing Technology Center Annual Conference and Exhibition (NECTEC-ACE 2011), Sep, 2011  Invited
  • KURABAYASHI Shuichi, KIYOKI Yasushi
    IEICE technical report, 111(77) 137-142, May 30, 2011  
    This paper proposes an impression-based social tagging system for video, which provides time-series color analysis mechanisms that extract aesthetic contexts from a video stream. This system provides a complex tag data structure that integrates an impression-based video analysis result and social collective intelligence for retrieving video data according to the emotional and mood preference of users. This system analyzes a video data by applying a psychological method of visual analysis to extract color-impression feature vectors corresponding to aesthetic and emotive contexts in the video. The system recommends users to assign a tag, which is a keyword or a term, to color-impression feature vectors for sharing the video analysis results. The unique feature of this system is an impression-based dynamic video navigation mechanism that makes the ranking of the videos by computing distance between the tags. When a user selects a tag, the system navigates users to other videos by ranking the videos according to the impression-based relevance to the selected tag. This impression-based navigation mechanism is effective to browse videos which are unfamiliar to users because users do not have to know detailed information. This impression-based social tagging system enable users to find videos by selecting tags according to their emotional requirements, such as "Find videos which are similar to the currently playing video in an overall impression" and "Find videos whose impression is similar to the currently playing video in local dramatic presentations".
  • KURABAYASHI Shuichi, KIYOKI Yasushi
    72 79-80, Mar 8, 2010  
  • SASAKI Shiori, KIYOKI Yasushi
    全国大会講演論文集, 72 297-298, Mar 8, 2010  
  • TAKAHASHI Yusuke, SASAKI Shiori, KIYOKI Yasushi
    IEICE technical report. Data engineering, 108(211) 27-28, Sep 14, 2008  
    This paper shows a design and a prototype implementation of "Futuris Text," the 4D world map system with semantic, temporal and spatial analyzers. This system enables defining multiple views of spatio-temporal and semantic relationships among historical documents to create various patterns of historical/cultural stories, according to user's viewpoints in a dynamic way. This system realizes "An Textbook with Computational Functions," which is descibed as an application of a meta-level database system that has functions of temporal, spatial and semantic computations.
  • SASAKI Shiori, KIYOKI Yasushi
    IEICE technical report. Data engineering, 108(211) 29-30, Sep 14, 2008  
    This paper presents a dynamic media-query creation method for semantic correlation computation between feature combinations and impressions of multiple media data sets. The main feature of our method is to create a media-query which reflects user's intention dynamically according to the feature-based combinations of media data selected by user's as context,and represents the distinctive impressions of media collection by time,culture,author,e.t.c.. The created media-query is very different from "query by example". We applied this query creation method to image retrieval and music retrieval to examine the feasibility of our method.
  • INOUE Katsuya, SASAKI Shiori, KIYOKI Yasushi
    IPSJ SIG Notes, 2007(65) 321-326, Jul 3, 2007  
    In this paper, we present an implementation method of a visualization system for realizing multiple views for historical documents. This system enables document classification by periodic information and visualize geographic information from simple plain texts about history. By using the implemented system by our method, users can grasp visually where historical events occurred or can speculate an era the processed document describe. To examine the availability of the system, we performed several experiments using documents on the free encyclopedia on the web which describe Japanese historical terms and events.
  • 高松 耕太, 倉林 修一, 佐々木 史織, 清木 康
    IPSJ SIG Notes, 2005(68) 531-538, Jul 14, 2005  
    In this paper, we present a context-based Pub/Sub system which enables a publication system to notify a user of incoming news articles of various domains according to the user's context automatically. To respond to diversification in informational consumption behavior, it is important to incorporate a context-dependent vector space model into Pub/Sub systems. However, there is at least one problem that the correlation values as a result of semantic computation vary depending on each domain of target documents. Therefore, the system needs some other automatic measures to decide what document should be notified, when it treats dynamic information resources such as daily-published-news. We present a method to detemine the context-dependent threshold automatically and adaptively by combining negative and positive sample documents. We also present an integrated query interpretation system with multiple vector spaces which reflect cross-cutting knowledge. We perform evaluation experiments and demonstrate the effectiveness our methods.
  • TAKAMATSU Kota, KURABAYASHI Shuichi, SASAKI Shiori, KIYOKI Yasushi
    IEICE technical report. Data engineering, 105(172) 205-210, Jul 7, 2005  
    In this paper, we present a context-based Pub/Sub system which enables a publication system to notify a user of incoming news articles of various domains according to the user's context automatically. To respond to diversification in informational consumption behavior, it is important to incorporate a context-dependent vector space model into Pub/Sub systems. However, there is at least one problem that the correlation values as a result of semantic computation vary depending on each domain of target documents. Therefore, the system needs some other automatic measures to decide what document should be notified, when it treats dynamic information resources such as daily-published-news. We present a method to detemine the context-dependent threshold automatically and adaptively by combining negative and positive sample documents. We also present an integrated query interpretation system with multiple vector spaces which reflect cross-cutting knowledge. We perform evaluation experiments and demonstrate the effectiveness our methods.
  • 大橋 英博, 清木 康, 石黒 晶子
    日本データベース学会letters, 2(3) 17-20, Dec, 2003  
  • IJICHI ASAKO, KIYOKI YASUSHI, SATO AKIRA, NAKAGAMI YASUHIRO
    IPSJ SIG Notes, 2003(71) 445-450, Jul 16, 2003  
    This paper presents a metadata generation method dealing with user's Viewpoints and its application to musical structures. In this paper, we use the term "Viewpoint" as the focussing point in listening to music. This method is userd for music-retrieval-by-impression according to user's Viewpoints. In general, music included various impression and its impression is fix only when the user's Viewpoint is specified. The main feature of this method is that Viewpoint is used for fixing impression of music metadata related to various user's Viewpoints. This paper shows several experimental results of the metadata creation to clarify the feasibility and effectiveness of our method.
  • IJICHI Asako, KIYOKI Yasushi, SATO Akira, NAKAGAMI Yasuhiro
    IEICE technical report. Data engineering, 103(191) 199-203, Jul 10, 2003  
    This paper presents a metadata generation mdhod dealing with user's Viewpoints and its application to musical structures. In this paper, we use the term "Viewpoint" as the focussing point in listening to music. This method is userd for music-retrieval-by-impression according to user's Viewpoints. In general, music included various impression and its impression is fix only when the user's Viewpoint is specified. The main feature of this method is that Viewpoint is used for fixing impression of music metadata related to various user's Viewpoints. This paper shows several experimental results of the metadata creation to clarify the feasibility and effectiveness of our method.
  • ISHIBASHI Naoki, KIYOKI Yasushi, NAKAGAMI Yasuhiro, SATO Akira
    IEICE technical report. Data engineering, 103(191) 265-270, Jul 10, 2003  
    For music data on the global area network, we propose an automatic extraction method of metadata that represent impression. The proposed method dynamically extracts adjectives and correlations according to each instrument of a datum, and it dynamically integrates the results of each instrument to generate the metadata of the datum. We clarify feasibility of the proposed method with an implemented system.
  • TAKAHASHI Yusuke, KIYOKI Yasushi
    IEICE technical report. Data engineering, 102(208) 121-126, Jul 11, 2002  
    In this paper we evaluate the effectiveness of a methodology for constructing a career-oriented learning-support. system on a multidatabase system. This methodology is used to integrate the lecture databases arid job definition databases in the multidatabase system architecture. This system enforces the support system for learning process at universities by providing a framework to construct learning plans. By applying this system, students are able to design their academic career according to their goals and the dynamic transitions of the society.
  • OMAE HIROKO, ISHIBASHI NAOKI, KIYOKI YASUSHI, ANZAI YUICHIRO
    IEICE technical report. Data engineering, 101(193) 143-150, Jul 12, 2001  
    With a rapid popularization of the global-area network, realizing an information retreaval method for media data is essentially important today. As forcusing to the information retriaval methods for music data, absolute methods to automatically extract impressions, which becomes metadata, of music data do not exist. In this paper, we propose an automatic metadata creation method for music with multiple musical instruments. The impression for each musical instrument is extracted from the music data which consist of multiple instruments and weighted according to each importance of the musical instrument. Our method extract the impression of a music datum by compounding the impression of each musical instrument with others. We clarity the effectiveness of the method in a view point of compounding the impression of each musical instrument by comparing to present methods, and the effectiveness of weighting by showing the result of the actual questionnaire.
  • ZUSHI Taizo, YOSHIDA Naofumi, KIYOKI Yasushi, KITAGAWA Takashi
    IPSJ SIG Notes, 2000(69) 331-338, Jul 26, 2000  
    In this paper we present a semantic knowledge discovery method using context dependent dynamic clustering for document data. The main feature of the method is to make clustering for document data semantically according to a given context. By using this method, we can dynamically obtain a set of semantic clusters of documents from a set of raw data according to a given context. We introduce a clustering method that further classifies each target cluster. This method is applied to the metadata attached to the document data formed in the semantic clustering. We clarify feasibility and effectiveness of the method by showing several experimental results using actual document data.
  • S Uemura, H Arisawa, M Arikawa, Y Kiyoki
    IEICE TRANSACTIONS ON INFORMATION AND SYSTEMS, E82D(1) 22-33, Jan, 1999  
    This paper surveys recent research activities on three major areas of digital media information base, namely, video database systems as a typical example of temporal application, database systems for mixed reality as an instance of spatial application, and kansei management for digital media retrieval as a case of humanistic feelings application. Current research results by the project Advanced Database Systems for Integration of Media and User Environments are reported.
  • SATO AKIRA, KIYOKI YASUSHI
    IEICE technical report. Data engineering, 97(160) 89-94, Jul 14, 1997  
    We have proposed a stream-oriented parallel processing scheme for creating multimedia data by integration continuous media data item. In this scheme, it is important to perform optimization for processor resource allocation in process for creating multimedia data. This paper presents an efficient computational method for obtaining optimal processor resource allocation for a query consisting of a set of basic operations in distributed processing environment.
  • HAYAMA Takanari, KIYOKI Yasushi
    IEICE technical report. Data engineering, 97(160) 83-88, Jul 14, 1997  
    While rendering multimedia documents, jitters and lags among media always araise as a common problem. Jitters and lags can be reduced by setting appropriate synchronization points among media. To set the synchronization points, duration of each medium is required. However, there are many multimedia documents with unpredictable duration such as user interaction which makes difficult to create static schedules of synchronizations. In this paper, we present a synchronization method for media whose duration is unpredictable. The synchronization method is based on coordinated computation.
  • HAYAMA Takanari, KIYOKI Yasushi
    IPSJ SIG Notes, 1996(45) 41-48, May 23, 1996  
    It is important to realize multimedia data processing for continuous media data in distributed computing environment. A sychronization mechanism among several contiunous media during the rendition is required for dynamic integration of those media. We designed and implemented a NAMI system which supports synchronizing processes to render continuous media data in the distributed computing environment. The NAMI system allows several independent applications to be arranged arbitrarily and work as they are integrated in a single application. Synchronization is supported by a Sync Manager of the NAMI system. The Sync Manager is included in each application. In this paper, the implementation method of the synchronization mechanism of the NAMI system is presented.
  • Muraoka Masanori, Sato Akira, Kiyoki Yasushi
    IPSJ SIG Notes, 1994(62) 225-232, Jul 20, 1994  
    We have proposed a stream-oriented parallel processing scheme which is applied to various application areas of databases. In this scheme, it is important to perform optimization for memory resource allocation in query processing for arbitrary combination of various database operations. This paper presents an efficient computational method for obtaining optimal memory resource allocation for a query consisting of an arbitrary set of database operations in distributed processing environments. Furthermore, this paper shows several experimental results for computations of actual memory resource allocation and discusses performance and effectiveness of the proposed method.
  • Sato Akira, Kiyoki Yasushi
    IEICE technical report. Data engineering, 93(251) 19-26, Sep 27, 1993  
    This paper presents an implementation method of a parallel processing system for supporting a wide variety of database applications.This system provides facilities to describe application-specific database operations and data structures in a functional programming way and to execute queries in parallel in the framework of functional computations. We have proposed the primitive set for parallel query processing. This primitive set is provided to realize the parallel processing system on arbitrary general purpose hardware environments.We can choose a hardware environment for this system reflecting requirements for the database application and advance of the hardware technology.
  • KIYOKI Yasushi
    IPSJ Magazine, 31(2) p225-233, Feb 15, 1990  

Books and Other Publications

 5

Presentations

 213

Teaching Experience

 34

Works

 4
  • 2001 - Present Others
    概要: 宇宙開発事業団における高度情報化、知識ベース化のための方式、システムおよびデータベースの構成方法についての検討を行ない、その指針を提示する。
  • 2000 - Present Others
    概要: データベースと感性情報についての最近の研究動向、および、我々の研究について、講演を行った。(〜平成13年まで)
  • 2000 - Present Others
    概要: マルチデータベースシステム、知識管理方式についての最近の研究(〜平成13年まで)
  • 1996 - Present Others
    概要: 日立製作所が開発したミドルウェアシステムOpenTP1に関する研究会の座長として、最近の分散システム、ネットワークシステムの最近の研究動向に関する講演、および、パネル討論の進行を行った。(〜平成12年まで)

Industrial Property Rights

 6