Abstract—Historical texts are an important component of cultural heritage, and are being digitized and published on the web in various portals for the researchers and the public. However, searching and linking them with related contents is challenging due to the non-structured text form, digitization errors, and the differences and variations between old and modern language, including historical names (e.g. places), used for querying. This paper addresses these issues by presenting an approach and a system for publishing old texts on the semantic web. As a case study, an existing historical newspaper archive on the web is considered. In our model, semantic metadata is added to the text using automated concept extraction methods. Search is implemented with semantic techniques, by creating a multi-faceted search interface for the text materials. Problems due to OCR errors and spelling variants are addressed with a fuzzy string matching algorithm trying to guess corresponding words in a lexicon, and giving suggestions for corrected word forms. References between texts in the library as well as links between the library and external knowledge sources are formed by using shared ontologies for semantic annotations. Keywords-historical newspapers; automatic semantic annotation; multi-faceted search; I.

Content in semantic web portals is often projected along application specific navigational taxonomies and linked semantically. This paper presents a logic-based method and a server ONTODELLA for these tasks. We argue that logic rules between the content layer and the application layer add flexibility and better architectural separation of content and functionality. The system has been implemented and applied succesfully in several semantic portals. 1

Searching and comparing information from semi-structured repositories is an important, but cognitively complex activity for internet users. The typical web interface displays a list of results as a textual list which is limited in helping the user compare or gain an overview of the results from a series of iterative queries. In this paper, we propose a new interactive, lightweight technique that uses multiple synchronized tag clouds to support iterative visual analysis and filtering of query results. Although tag clouds are frequently available in web interfaces, they are typically used for providing an overview of key terms in a set of results, but thus far have not been used for presenting semi-structured information to support iterative queries. We evaluated our proposed design in a user study that presents typical search and comparison scenarios to users trying to understand heterogeneous clinical trials from a leading repository of scientific information. The study gave us valuable insights regarding the challenges that semi-structured data collections pose, and indicated that our design may ease cognitively demanding browsing activities of semi-structured information.

The LogosFlow system used by the District Six museum to capture and browse artefact collections is neither userfriendly nor intuitive. It was decided that an entirely new system be built and a user-centered design approach be taken to achieve this. Meetings were held with the collections staff, problems with the previous system identified and possible solutions brainstormed. Humancomputer interaction methodologies were applied to the user interfaces and the usability of the resulting interfaces was subsequently tested by the staff. It was found that the proposed system was more intuitive and much easier to use than the LogosFlow system and that with the addition of minor extensions it might be a suitable replacement for the current system.

The Flamenco system [7, 6] is a web search interface that allows users to browse through large data sets using predefined hierarchical faceted metadata. It is built on top of a conventional relational database and currently scales to collections of several tens of thousands of items. In the current implementation, the system translates each user query into multiple SQL group-by commands in order to obtain query preview information for possible future queries. These group-by's take up a significant fraction of the query processing time. In this note, we describe an optimization that allows us to speed up the group-by computations dramatically. Our ideas have some similarity to the work of Beyer and Ramakrishnan on computing iceberg data cubes [4].

who was the first Ph.D. I knew. iii Acknowledgements This thesis is the product of both hard work and invaluable assistance from colleagues and friends. In my time at Penn, I have had the privilege of working with some of the finest faculty and students. First and foremost, I thank my adviser Professor Fernando Pereira, who has expertly guided my research and career. Much can be said about his amazing breadth and depth of knowledge in so many fields. Whenever I mention his name to colleagues in the field, I always hear praise of his kindness and character. This has made him a wonderful adviser and mentor. Next, I thank my committee, who have shaped both my thesis and my career in the field. Professor Mitch Marcus and I have shared many long and fascinating conversations, many that had nothing to do with computer science. I will miss our frequent discussions and look forward to having more in the future. Professor Ani Nenkova read my thesis with a careful eye and provided invaluable critical feedback. Our conversations gave me

Is searching creative? Searching and information seeking are part of the creative process. An architect looking for “seed ” ideas for a new project may search an architecture database. Novelists, journalists and artists may similarly search the web for new ideas. A historian will explore archival material for a research project. Even graduate students may employ search as they refine and narrow their research

During my doctoral research I investigate how visual exploration affects information seeking on the Web. I am particularly interested in the role of interactive visualization in exploratory search interfaces, development of web-based systems, and new ways of information seeking. In this research overview, I provide an introduction to my research topic, outline prior related work, discuss my methodology and challenges, and present a timeline for my research progress. 1