Showing papers by "Katsumi Tanaka published in 2004"

Temporal and spatial attribute extraction from web documents and time-specific regional web search system

Abstract: Regional web search engines provide users with spatially restricted web contents. We propose a time-specific regional web search engine, where users can retrieve both spatially and temporally restricted information from the Web, i.e. a description of a shop available at certain time in a certain region. Because there are many web pages containing descriptions of more than one geographic entity, a matching mechanism of spatial and temporal attributes is required. We used HTML tag structures to make matches between addresses and temporal interval expressions. Based on the extracted result, map-based visual interface was implemented.

Extraction of cognitively-significant place names and regions from web-based physical proximity Co-occurrences

Abstract: The cognitive significances of geographic objects play an important role in optimizing a set of place names presented on a map interface. The aim of this research is to assign a value for cognitive significance to each place name in GIS, based on regional documents collected from the Web. The assigned values could be used in GIS applications including regional information searches, path findings, and car navigation systems. We will compare different criteria for cognitive significance and discuss a system already in use, before presenting our experimental results.

Complementary information retrieval for cross-media news content

Abstract: In this paper, we propose a new way of integrating cross-media news content, such as television programs and web pages. We search cross-media news content to find complementary items which can provide additional information to users interested in a particular topic. The complementary news items searched for are not just similar to the item the user is interested in, but also provide information in more detail or from a different perspective. First, we propose a novel content representation model called the "topic structure" model. Intuitively, a topic structure is made up of a pair of subject and content terms. Subject terms denote the dominant terms of a news item. A content term is a term having strong co-occurrence relationships with the subject terms. Based on the topic structure, we search for information related to a given news item (e. g. , one in which the user is interested) from content, context, and media complementation perspectives. We also describe an application system which concurrently presents a television news program along with complementary news articles to help users understand news topics in greater detail and from multiple perspectives.

Topic-Structure Based Complementary Information Retrieval for Information Augmentation

Abstract: It becomes possible to acquire our interesting information from diverse information source and media to improve the quality and detailedness of information. For instance, with the spreading of digital broadcasting and broadband internet connection services, the infrastructure for integration of TV-programs and the Internet is prepared and we can find the additional information of a TV-program from the Web. In this paper, we propose a novel information retrieval mechanism for information augmentation based on the notion of topic structure. One of the notable features of our retrieval mechanism is that it can be used to find complementary information of a given web page or video. That is to say, the retrieved information is not just similar to the given web page or video. It can provide some additional information to detail the given one or describe it from different perspective. In this paper, we also show some evaluation results of the complementary information retrieval mechanism.

Tv2web: generating and browsing web with multiple lod from video streams and their metadata

Abstract: We propose a method of automatically constructing Web content from video streams with metadata that we call TV2Web. The Web content includes thumbnails of video units and caption data generated from metadata. Users can watch TV ona normal Web browser. They can also manipulate Web content with zooming metaphors to seamlessly alter the level of detail (LOD) of the content being viewed. They can search for favorite scenes faster than with analog video equipment, and experience a new cross-media environment. We also developed a prototype of the TV2Web system and discuss its implementation.

Query Modification by Discovering Topics from Web Page Structures

Abstract: We propose a method that identifies from Web pages pairs of keywords in which one word describes the other and uses these relations to modify the query. It takes into account the positions of the words in the page structures when counting their occurrences and applies statistical tests to examine the differences between word co-occurrence rates. It finds related keywords more robustly regardless of the word type than the conventional methods, which do not consider page structures. It can also identify subject and description keywords in the user’s input and find additional keywords for detailing the query. By considering the document structures, our method can construct queries that are more focused on the user’s topic of interest.

Retrieving regional information from web by contents localness and user location

Abstract: As use of the World Wide Web (WWW) and mobile devices spreads and becomes more advanced, people are increasingly able to use mobile devices, such as cell phones, to access various kinds of local information through the Internet for use in daily activities. Conventional mobile information services focus on providing location information related to the user's current location. In contrast, we propose an information retrieval method that searches for 'local information' pertaining to a wider geographic region. In this paper, we consider local information to be the kind of information that is of interest only to users within certain regions or organizations. From this perspective, 'local information' differs from location information. Moreover, we are interested in searching for local information pertaining to a wider region (an area) that extends beyond a specific location (a place). Our local information retrieval method includes three stages: 1) searching for information using a query generated with user location data and keywords, 2) mapping the content coverage of searched Web pages onto the user's current region, and 3) ranking the regional information based on the notion of localness degree.

Discovering aspect-based correlation of Web contents for cross-media information retrieval

Abstract: The main issue regarding cross-media information retrieval is the determination of correlations between different types of media objects. The conventional approach derives the correlations based on common properties extracted from media contents or synchronous presentation of multiple media pre-authored in a scheduled scenario. We propose a novel approach for determining the cross-media correlation derived from the referential contexts of media objects in the Web. A Web page links to the media objects distributed over the Web so that it aggregates them with respect to the page content. Our approach extracts the referential context by analyzing the logical structure of the Web and discovers the aspect of a media object, which means the latent semantics of the referential context. The aspect-based correlation reveals the relation between media objects regarding their reputations on the Web. In this paper, we propose an approach for discovering aspect-based correlations with an experimental implementation

Relative queries and the relative cluster-mapping method

Abstract: Most conventional database systems and information retrieval systems force users to specify the qualification conditions in queries in an absolute manner. That is, a user must specify a qualification condition to be satisfied by each retrieval result. To do this properly, users must have sufficient knowledge about the metadata and their structures. In the real world, however, people often specify a qualification in a relative manner such as I prefer this one among these. In this paper, we propose the notion of relative queries, and their query processing method called the relative cluster-mapping. A relative query specifies user's selection criteria as selecting his/her favorite data among a sample data cluster. This notion is useful when users do not have sufficient knowledge about metadata, or users cannot decide a complete qualification condition. The relative cluster-mapping method maps the relative position of the user-selected data in a sample data cluster to a target data cluster and returns an answer from the target data cluster. The answer's relative position in the target data cluster is similar to that of the selected data in a sample data cluster. We believe it is more natural to express the desired data using relative qualifications. We also have developed prototype system, and evaluated its appropriateness.

Relative Queries and the Relative Cluster-Mapping Method

Abstract: Most conventional database systems and information retrieval systems force users to specify the qualification conditions in queries in an ”absolute” manner. That is, a user must specify a qualification condition to be satisfied by each retrieval result. To do this properly, users must have sufficient knowledge about the metadata and their structures. In the real world, however, people often specify a qualification in a relative manner such as ”I prefer this one among these.” In this paper, we propose the notion of ”relative queries,” and their query processing method called the ”relative cluster-mapping.” A relative query specifies user’s selection criteria as selecting his/her favorite data among a sample data cluster. This notion is useful when users do not have sufficient knowledge about metadata, or users cannot decide a complete qualification condition. The ”relative cluster-mapping” method maps the relative position of the user-selected data in a sample data cluster to a target data cluster and returns an answer from the target data cluster. The answer‘s relative position in the target data cluster is similar to that of the selected data in a sample data cluster. We believe it is more natural to express the desired data using relative qualifications. We also have developed prototype system, and evaluated its appropriateness.

A3: framework for user adaptation using xslt

Abstract: We propose a system called "Adaptation Anywhere & Anytime(A3)", which is a framework for making web sites/applications adaptable to user's needs or interests, and we describe the implement of a web site on A3 by using XSLT. Web sites/applications built on A3 construct user ontologies for each user automatically and share them between sites/applications. Each site/application uses the user ontology to select an appropriate resource for the user and to present such resources in a suitable form. And A3 offers the method for constructing the adaptable web sites using XSLT. The author of web sites can easily make their sites adaptable by using XSLT.

Retrieving Relevant Portions from Structured Digital Documents

Abstract: Retrieving relevant portions from structured documents consisting of logical components has been a challenging task in both the database and the information retrieval world, since an answer related to a query may be split across multiple components. In this paper, we propose a query mechanism that applies database style query evaluation in response to IR style keyword-based queries for retrieving relevant answers from a logically structured document. We first define an appropriate semantics of keywords-based queries and then propose an algebra that is capable of computing every relevant portion of a document, which can be considered answer to a set of arbitrary keywords. The ordering and structural relationship among the components are preserved in the answer. We also introduce several practically useful filters that saves users from having to deal with an overwhelming number of answers.

Discovering Aspects of Web Pages from Their Referential Contexts in the Web

Abstract: There are an enormous number of Web pages of unknown authorship, and even though Web search engines precisely evaluate the relevancy of Web page contents, a user cannot be sure whether a search result shows credible information. Considering that a Web page is referred to by other pages in various contexts through links, these contexts indicate the reputation of the page. For example, some pages may refer to a company’s page as “an excellent local company” and still other pages may refer to it as “a member of a certain research project”, while the company’s page itself might contain only product and service information. Such references are called “aspects” of the Web page, as distinguished from the content of the page. In this paper, we propose an approach for discovering aspects for characterizing Web pages based on their contexts. We define criteria for selecting “aspectual” Web content based on (1) its strength of association with the page based on the logical structure of the Web (i.e. Web document structure and link structure), (2) its novelty of content compared to the page and (3) its typicality among multiple contexts. We evaluate how these criteria affect aspect discovery results. We also explain the grouping of Web pages based on aspect similarity. This helps us to find Web pages that are referred to in the same way even though their content is different.

Discovering aspects of Web pages from their referential contexts in the Web

Abstract: There are an enormous number of Web pages of unknown authorship, and even though Web search engines precisely evaluate the relevancy of Web page contents, a user cannot be sure whether a search result shows credible information. Considering that a Web page is referred to by other pages in various contexts through links, these contexts indicate the reputation of the page. For example, some pages may refer to a company's page as an excellent local company and still other pages may refer to it as a member of a certain research project, while the company's page itself might contain only product and service information. Such references are called aspects of the Web page, as distinguished from the content of the page. In this paper, we propose an approach for discovering aspects for characterizing Web pages based on their contexts. We define criteria for selecting aspectual Web content based on (1) its strength of association with the page based on the logical structure of the Web (i.e. Web document structure and link structure), (2) its novelty of content compared to the page and (3) its typicality among multiple contexts. We evaluate how these criteria affect aspect discovery results. We also explain the grouping of Web pages based on aspect similarity. This helps us to find Web pages that are referred to in the same way even though their content is different.

How to build an adaptive web site: A Framework for User Adaptation

Abstract: We propose a framework for a user adaptation, called Adaptation Anywhere & Anytime(“A3”). User profiles for a user adaptation are automatically constructed and shared between web sites on A3. We also provide web-sites authors with a simple method for building an adaptive web site on A3 using XSLT.

My portal viewer for content fusion based on user's preferences

Abstract: A novel Web application called "my portal viewer (MPV)" has been developed to provide Web users with higher quality content, which is needed due to a rapidly growing amount of content on the Web. It provides fused news to the user, based on two viewpoints, through a user friendly interface and the user's preferences. MPV automatically selects and merges content from many news pages, based on the user's interest and knowledge, after gathering these pages from various news Web sites. Our unique approach is that the layout of the MPV page is applied to the users' favorite news portal page and a part of the original content is replaced by the fused content. Whenever a user accesses an MPV page after browsing other news pages, he/she can acquire the desired content efficiently because MPV presents a refreshed page, based on the user's behavior, which reflects his/her interests and knowledge. In addition to the MPV framework, methods that are based on user reference for replacing and selecting have been developed using an HTML table model.

Time-based contextualized-news browser (t-cnb)

Abstract: We propose a new way of browsing contextualized-news articles. Our prototype browser system is called a Time-based Contextualized-News Browser (T-CNB). The T-CNB concurrently and automatically presents a series of related pages for one news source while browsing the user-specified page. It extracts the past related pages from a user-specified news articles on the web. The related pages outline the progress of user-specified news articles. We call the related pages 'contextual pages'.Using the T-CNB, a user only needs to specify one news article on the web. The user then automatically receives past related news articles, which provide a wider understanding of the topic. The T-CNB automatically generates and presents contextualized news articles.

Aspect Discovery: Web Contents Characterization by Their Referential Contexts

Abstract: A web page is referred to by other pages through links in various contexts, and these contexts indicate the “customer’s viewpoint” for the page. The references are called “aspects” of a web page, as distinguished from the content of the page. In this paper, we propose an approach for discovering aspects to characterize web pages based on their referential context. Based on the logical structure of the web (i.e., the web document structure and link structure), our approach discovers the appropriate range of surrounding contents and assigns them as the context of the web page. The aspects of the web page are discovered by clustering multiple contexts so that each aspect represents a “typical reference” to the page. The aspect can be used to strengthen the usability and credibility of a web page.

Device Cooperative web Browsing and retrieving mechanism on ubiquitous networks

Abstract: We propose a new concept for displaying, browsing, and retrieving web content in public spaces via a ubiquitous network. The new type of content operation, described in this paper, is semantically synchronized and based on Device Cooperative Browsing and Device Cooperative Retrieval. These are interactive methods of browsing and retrieving content related to that currently being viewed by the user. The operating mechanism is based on multiple devices connected via a peer-to-peer network. The mechanism operates cooperatively to find each user's individual interests and it maintains information about them. Multiple devices sharing common information might thus have common topics of interest listed on various user devices. When new devices are connected in this environment, they detect the required information through filtering, and obtain it from other devices in the local ubiquitous network. In addition to describing this concept, we also discuss the WebBoard, which is a successful implementation of our novel approach.

Facilitating understanding for children by translating web contents into a storybook

Abstract: This paper describes a medium, called Interactive e-Hon, for helping children to understand contents from the Web. It works by transforming electronic contents into an easily understandable “storybook world”. In this world, easy-to-understand contents are created generated by creating 3D animations that include contents and metaphors, and by using a child-parent model with dialogue expression and a question-answering style comprehensible to children.

Query relaxation and answer integration for cross-media meta-searches

Abstract: The amount of information available on the Web has increased dramatically. This information comes in a wide variety of media types. Therefore, to retrieve information on a topic using conventional search engines, users must search many sites from several different aspects. Thus, it is important to integrate and organize this information from the different sites. We propose an integration and organization system based on a query relaxation approach for cross-media meta-search engines. Frequently, the parameters for information retrieval are too specific or exacting to generate any relevant sites. However, by gradually relaxing the search terms used for information retrieval, we can solve this problem while narrowing the search to sites that are most relevant to the subject being researched. We show several examples of how the relaxation approach works as well as ways in which it can be applied. We also demonstrate the advantages of our approach and future work for this research.

Extraction of cognitively-significant place names and regions from Web-based physical proximity co-occurrences

Abstract: The cognitive significances of geographic objects play an important role in optimizing a set of place names presented on a map interface. The aim of this research is to assign a value for cognitive significance to each place name in GIS, based on regional documents collected from the Web. The assigned values could be used in GIS applications including regional information searches, path findings, and car navigation systems. We will compare different criteria for cognitive significance and discuss a system already in use, before presenting our experimental results.

Device Cooperative Web Browsing and Retrieving Mechanism on Ubiquitous Networks

Abstract: We propose a new concept for displaying, browsing, and retrieving web content in public spaces via a ubiquitous network. The new type of content operation, described in this paper, is semantically synchronized and based on Device Cooperative Browsing and Device Cooperative Retrieval. These are interactive methods of browsing and retrieving content related to that currently being viewed by the user. The operating mechanism is based on multiple devices connected via a peer-to-peer network. The mechanism operates cooperatively to find each user’s individual interests and it maintains information about them. Multiple devices sharing common information might thus have common topics of interest listed on various user devices. When new devices are connected in this environment, they detect the required information through filtering, and obtain it from other devices in the local ubiquitous network. In addition to describing this concept, we also discuss the WebBoard, which is a successful implementation of our novel approach.

Visualization and Sharing of Web History on Group-Based Web Browsing

Abstract: Recently, information retrieval on Web browsing is not always for individual use. It is considered that demands for the information presentation to other users has been increasing. It is useful to get to know the back- ground as which the contents recommended by other users were chosen in the environment where a group shares the contents obtained by individual. Because if users know what others think to search , how many similar pages to recommendation they browse or what range they peruse the Web contents ,users can know whether recommendation contents is selected carefully and perused comprehensively. In this paper, we propose a way to share selectiveness

Continuous web: a new image-based hypermedia and scape-oriented browsing

Abstract: Conventionally, Web pages have been recognized as documents described by HTML. Image data, such as photographs, logos, maps, illustrations, and decorated text, have been treated as sub-components of Web documents. However, we can alternatively recognize all Web pages as images on the screen. When a Web page is treated as an image, its HTML data is considered to be metadata which describes the image content. Taking such a viewpoint, we propose a new image-based hypermedia which we call continuous web. In our model, there is no distinction between Web images and other images such as photographs.Regarding everything on the Web as images leads us to consider a new style of browsing and navigating. We use the term scape-oriented browsing. We define a scape as a collection of continuously accumulated images. For example, whenever we walk in the real world, we can perceive and remember various forms of information through a scape process. Here, we describe new methods for scape-oriented browsing, such as see-through anchors, parallel navigation, and peripheral scape presentation. We have designed and implemented a prototype system based on our model. Our system offers continuous browsing and navigation to users. We explain our concepts and discuss the effectiveness and potential of this approach.

TV2Web: generating and browsing Web with multiple LOD from video streams and their metadata

Abstract: We propose a method of automatically constructing Web content from video streams with metadata that we call TV2Web. The Web content includes thumbnails of video units and caption data generated from metadata. Users can watch TV on a normal Web browser. They can also manipulate Web content with zooming metaphors to seamlessly alter the level of detail (LOD) of the content being viewed. They can search for favorite scenes faster than with analog video equipment, and experience a new cross-media environment. We also developed a prototype of the TV2Web system and discuss its implementation. Our experimental result shows 35% improvement for scene search tasks of video streams.

Query modification by discovering topics from Web page structures

Abstract: We propose a method that identifies from Web pages pairs of keywords in which one word describes the other and uses these relations to modify the query. It takes into account the positions of the words in the page structures when counting their occurrences and applies statistical tests to examine the differences between word co-occurrence rates. It finds related keywords more robustly regardless of the word type than the conventional methods, which do not consider page structures. It can also identify subject and description keywords in the user's input and find additional keywords for detailing the query. By considering the document structures, our method can construct queries that are more focused on the user's topic of interest.

Guiding Web Search by Third-party Viewpoints: Browsing Retrieval Results by Referential Contexts in Web

Abstract: In a typical web search session, a user starts with partially defined queries which are refined as the user finds out more about the search domain. Conventional approaches emphasize narrowing the search, but pay little attention to outlining the scope of search, in which the user aggregates the retrieval results in a bottom-up approach. Although the standard technique is to cluster the page contents, this becomes difficult as the results contain a wide spectrum of page contents due to the fragmentary queries. In this paper, we propose an approach for outlining the scope of search based on “third-party viewpoints” on retrieval results. A web page is referred to by other web pages in various contexts through hyperlinks. The proposed method extracts an appropriate range of web contents surrounding the page as the referential context. By clustering similar contexts, web page “aspects” are discovered as generalized descriptions of the page references. Visualizing correlations between the page contents and the aspects helps a user find coherent semantic units of the retrieval results, which are classified into acknowledged pages, versatile pages, and alternative pages. We explain the details of the proposed method and a prototype implementation.

Guiding Web search by third-party viewpoints: Browsing retrieval results by referential contexts in Web

Abstract: In a typical web search session, a user starts with partially defined queries which are refined as the user finds out more about the search domain. Conventional approaches emphasize narrowing the search, but pay little attention to outlining the scope of search, in which the user aggregates the retrieval results in a bottom-up approach. Although the standard technique is to cluster the page contents, this becomes difficult as the results contain a wide spectrum of page contents due to the fragmentary queries. In this paper, we propose an approach for outlining the scope of search based on third-party viewpoints on retrieval results. A web page is referred to by other web pages in various contexts through hyperlinks. The proposed method extracts an appropriate range of web contents surrounding the page as the referential context. By clustering similar contexts, web page aspects are discovered as generalized descriptions of the page references. Visualizing correlations between the page contents and the aspects helps a user find coherent semantic units of the retrieval results, which are classified into acknowledged pages, versatile pages, and alternative pages. We explain the details of the proposed method and a prototype implementation.

Topic-structure based complementary information retrieval for information augmentation

Abstract: It becomes possible to acquire our interesting information from diverse information source and media to improve the quality and detailedness of information. For instance, with the spreading of digital broadcasting and broadband internet connection services, the infrastructure for integration of TV-programs and the Internet is prepared and we can find the additional information of a TV-program from the Web. In this paper, we propose a novel information retrieval mechanism for information augmentation based on the notion of topic structure. One of the notable features of our retrieval mechanism is that it can be used to find complementary information of a given web page or video. That is to say, the retrieved information is not just similar to the given web page or video. It can provide some additional information to detail the given one or describe it from different perspective. In this paper, we also show some evaluation results of the complementary information retrieval mechanism.