Search

Bias in the retrieval of documents can directly influence the information access of a digital library. In the worst case, systematic favoritism for a certain type of document can render other parts of the collection invisible to users. This potential bias can be evaluated by measuring the retrievability for all documents in a collection. Previous evaluations have been performed on TREC collections using simulated query sets. The question remains, however, how representative this approach is of more realistic settings. To address this question, we investigate the effectiveness of the retrievability measure using a large digitized newspaper corpus, featuring two characteristics that distinguishes our experiments from previous studies: (1) compared to TREC collections, our collection contains noise originating from OCR processing, historical spelling and use of language; and (2) instead of simulated queries, the collection comes with real user query logs including click data. First, we assess the retrievability bias imposed on the newspaper collection by different IR models. We assess the retrievability measure and confirm its ability to capture the retrievability bias in our setup. Second, we show how simulated queries differ from real user queries regarding term frequency and prevalence of named entities, and how this affects the retrievability results.

This deliverable presents the initial Reference Architecture of the VRE4EIC Project. The Reference Architecture is one of the results of the first year of work of the project. In order to derive it, the project conducted more than sixty interviews, and carefully characterized five existing e-Research Infrastructures. This way, it obtained a significant picture of the needs of the scientists, and of how these needs are met by the existing e-Research Infrastructures. This picture was an input to the architects of the Project, who analysed it in detail, and distilled an exhaustive list of functions whose implementation is needed to satisfy the requirements in ways that advance the current e-Research Infrastructures. Competence in software system design and development, as well as consideration of existing standards and best practices, allowed the VRE4EIC architects to group functions into components, and to further structure components in sub-components, thereby laying the basic building blocks of the Reference Architecture. Each component or sub-component was assigned a set of interfaces, each consisting of methods performing the tasks that provide the lowest level of granularity of the Reference Architecture. An intense activity of verification then ensued, aiming at verifying that the methods on the component interfaces could be appropriately composed into complex workflows able to realize the required functionality. The deliverable presents the result of this work, and documents the activities and the decisions that led to it. Based on this solid piece of work, the project is entering its second year of activity. While keeping the Reference Architecture aligned with additionally collected requirements and e-RI characterizations, a gap analysis will be conducted to assess two existing Virtual Research Environments with respect to the Reference Architecture. The gap analysis will allow the project to select the components of the Reference Architecture that will be implemented in order to enhance the selected Virtual Research Environments. As a result of these activities, at the end of the project the final Reference Architecture will be released.

The ever growing number of textual historical collections calls for methods that can meaningfully connect and explore these. Different collections offer different perspectives, expressing views at the time of writing or even a subjective view of the author. We propose to connect heterogeneous digital collections through temporal references found in documents as well as their textual content. We evaluate our approach and find that it works very well on digital-native collections. Digitized collections pose interesting challenges and with improved preprocessing our approach performs well. We introduce a novel search interface to explore and analyze the connected collections that highlights different perspectives and requires little domain knowledge. In our approach, perspectives are expressed as complex queries. Our approach supports humanity scholars in exploring collections in a novel way and allows for digital collections to be more accessible by adding new connections and new means to access collections.

We provide a detailed report of a reproduction study of a paper published in the International Journal of Medical Sciences (IJMS). We first use the PROV-O ontology to model our reconstruction of the computational workflow of the original experiment and to systematically explicate all information that is needed for an reproduction study. We then identify which part of the required information is published in the IJMS paper and what part is missing. We then discuss our reproduction of this workflow, following the original as much as possible. Again, we use PROV-O to precisely define our version of the workflow, including our version of the information that was missing in the IJMS paper of the study. Finally, we generalize from the specific cased described in the original paper by providing a web service that allows mining for arbitrary drug-adverse event pairs.

Enterprises are experiencing the need to model their domain in order to improve the communication and control over their intellectual assets. There are several different approaches to modeling a domain, some old-fashioned, some trendy; practitioners are at a loss when having to decide which approach best suits their situation. In this research we compare different domain modeling schemes to find out which problem each of them is best suited to solve. The modeling schemes we consider are taxonomy, thesaurus and ontology. We draw relevant information from evaluating case studies and interviewing practitioners in the field of domain modeling. We restrict the scope to those projects that aim at improving the retrieval of information. The aim of this research is to support practitioners when they want to model their domain to choose the right scheme, considering the trade-offs being made between requirements and effort.

Pl@ntNet is an innovative participatory sensing platform relying on image-based plants identification as a mean to enlist non-expert contributors and facilitate the production of botanical observation data. One year after the public launch of the mobile application, we carry out a self-critical evaluation of the experience with regard to the requirements of a sustainable and effective ecological surveillance tool. We first demonstrate the attractiveness of the developed multimedia system (with more than 90K end-users) and the nice self-improving capacities of the whole collaborative workflow. We then point out the current limitations of the approach towards producing timely and accurate distribution maps of plants at a very large scale. We discuss in particular two main issues: the bias and the incompleteness of the produced data. We finally open new perspectives and describe upcoming realizations towards bridging these gaps.

Door de explosie van informatie op het web gebruiken meer mensen dan ooit online-zoekapplicaties om antwoorden te vinden in stapels digitale documenten. Door deze ontwikkeling wordt onderzoek naar zoekgedrag en naar hoe mensen gebruikmaken van zoekapplicaties steeds belangrijker. Alia Amin onderzocht zoektaken in meerdere bronnen, zoals het internet, databases en digitale bibliotheken, alsmede de invloed van verschillende typen gebruikers en domeinen op zoekgedrag. Het onderzoek biedt inzicht in de zoektaken van verschillenden typen gebruikers en domeinen. Amin bestudeerde het zoekgedrag van domeindeskundigen in het cultureel erfgoed en het zoekgedrag van onervaren gebruikers op het gebied van mobiel zoeken. Daarnaast verrichtte zij onderzoek naar het ontwerpproces en evaluatie van nieuwe, innovatieve interfaces om specifieke zoektaken te ondersteunen. De resultaten uit het onderzoek bieden een aantal wegwijzers voor toekomstig onderzoek in dit multidisciplinaire veld. Amin identificeerde eisen om complexe zoektaken in meerdere bronnen te ondersteunen. Bovendien ontwikkelde ze een aantal applicaties als voorbeeld. Deze applicaties kunnen dienen als inspiratie voor zowel de Semantic Web-gemeenschap als de Human Computer Interaction-gemeenschap. Amin verrichtte haar onderzoek aan het Centrum Wiskunde & Informatica in Amsterdam.