• DARIAH EU
• Open Access close
• Hyper Article en Ligne close
• ISTI Open Portal close

International audience; The digital text platform is part of the Flemish contribution to DARIAH Belgium (DARIAH = Digital Research Infrastructure for the Arts and Humanities). The goal is to create a platform for the collaborative management and discovery of digitised textual collections that allows digital humanities researchers to prepare their corpora (consisting of, for example, digitised newspapers and books) for textual analysis. The platform will enable researchers to browse and search the digitised collections compiled, cleaned, enriched and managed by the researchers themselves. Once the relevant research sub-corpus has been compiled, data export tools, using standardised open formats (such as XML, JSON, .csv, .txt, etc.) will enable researchers to export sub-corpus for analysis with existing digital text analysis tools such as MALLET, (http://mallet.cs.umass.edu/topics.php) for topic modelling, VOYANT (http://voyant-tools.org) for data visualisation or AntConC (http://www.laurenceanthony.net/software/antconc/) for concordance and textual analysis.The platform has been conceived as part of a larger and modular virtual research environment service infrastructure (http://www.ghentcdh.ugent.be/projects/dariah-vl_vre.si). In a previous phase, possible frameworks and content management systems were tested, notably Islandora (a digital asset management system based on Fedora Commons and Drupal), but also Mediawiki and Omeka.One of the main challenges of the envisaged new platform is the possibility to integrate a wider variety of possible textual data streams (including a scan workflow). In addition, user-friendliness, scalability, adherence to standards and facilitating the interoperability of data are key issues to be addressed. The platform will build on the existing IIIF format, the International Image Interoperability Framework. This format is used by some of the most important libraries and cultural heritage institutions in the world, therefore providing access to enormous collections of digital objects. As the name suggests, IIIF is mainly focused on displaying and annotating images. However, we fully endorse the IIIF-community’s vision to develop an overarching interoperability framework for other data types, including all kinds of textual data. Benefits of the format include the interoperability, the ease of sharing images and annotations without the need to exchange files, and its support for multilingual data. In the months leading up to the conference, we will evaluate the existing IIIFpowered digital libraries and research projects and how they deal with practices of co-creation, data cleaning and enrichment of (structural) metadata. OCR improvement will become vital, as digital textual analysis can only be performed well on high-quality textual data. A related challenge will be combining the various input formats and converting them to different output formats required for analysis. In our poster, we will present a summary of our experiences with and technical assessment of our previous Islandora installation, in addition to our survey of the existing corpus management solutions. As a way of conclusion, we will introduce the envisioned new version of the platform.

This article shows that metadata plays a central role in our society and concludes that through collaborative work, it is possible to pool solutions and to establish relationships of cooperation, both at the level of practical tool development and with regard to sharing and creating knowledge and know-how. ispartof: ABB: Archives et Bibliothèques de Belgique - Archief- en Bibliotheekwezen in België vol:106 pages:135-144 status: published

Obtaining high-quality labelled data for training a classifier in a new application domain is often costly. Transfer Learning (a.k.a. “Inductive Transfer”) tries to alleviate these costs by transferring, to the “target” domain of interest, knowledge available from a different “source” domain. In transfer learning the lack of labelled information from the target domain is compensated by the availability at training time of a set of unlabelled examples from the target distribution. Transductive Transfer Learning denotes the transfer learning setting in which the only set of target documents that we are interested in classifying is known and available at training time. Although this definition is indeed in line with Vapnik’s original definition of “transduction”, current terminology in the field is confused. In this article, we discuss how the term “transduction” has been misused in the transfer learning literature, and propose a clarification consistent with the original characterization of this term given by Vapnik. We go on to observe that the above terminology misuse has brought about misleading experimental comparisons, with inductive transfer learning methods that have been incorrectly compared with transductive transfer learning methods. We then, give empirical evidence that the difference in performance between the inductive version and the transductive version of a transfer learning method can indeed be statistically significant (i.e., that knowing at training time the only data one needs to classify indeed gives an advantage). Our clarification allows a reassessment of the field, and of the relative merits of the major, state-of-the-art algorithms for transfer learning in text classification.

International audience

The European Holocaust Research Infrastructure (EHRI) portal website aims to aggregate digitally available archival descriptions concerning the Holocaust. This portal is actually a meta-catalogue, or an information aggregator, whose biggest goal is to have up-to-date information by means of building sustainable data pipelines between EHRI and its content providers. Just like in similar archival information aggregators (e.g. Archives Portal Europe or Monasterium), the XML-based metadata standard Encoded Archival Description (EAD) plays a key role. The article presents how EADs are imported into the portal, mainly thanks to the Open Archive Initiative protocols.

In this paper we describe the development and evaluation of a visual analytics tool to support historical research. Historians continuously gather data related to their scholarly research from archival visits and background search. Organising and making sense of all this data can be challenging as many historians continue to rely on analog or basic digital tools. We built an integrated note-taking environment for historians which unifies a set of func-tionalities we identified as important for historical research including editing, tagging, searching, sharing and visualization. Our approach was to involve users from the initial stage of brainstorming and requirement analysis through to design, implementation and evaluation. We report on the process and results of our work, and conclude by reflecting on our own experience in conducting user-centered visual analytics design for digital humanities. International audience

International audience; This article considers need for more visible, available, accessible, innovative and shared bibliographic data in the internet age and the subsequent benefits of these transformations for galleries, libraries, archives and museums. Recent and ongoing research and development activities in the following fields are explored: entity identification, reconciliation, data enrichment, MARC records enriched with URIs, conversion to RDF, creation of relationship criteria for the improved identification of entities and a knowledge base of clusters that uses the paradigms of the semantic web. These improvements are discussed in the context of the BIBFRAME (Bibliographic Framework Initiative) data model and associated projects such as SHARE-VDE. The aim of the article is to outline current and future research and development activities in collaboration with the library community concerning the dissemination and discoverability of bibliographic data and research knowledge.

Contemporary researchers are beginning to explore the possibilities and opportunities of digital humanities, but encounter major obstacles regarding (meta)data1. Many archival institutions lack the technology or the skills to process (meta)data, let alone share it. Different types of standards coexist and sometimes conflict with each other, while implementing the same standards often leads to slight differences which decrease interoperability. Storing, securing and making born-digital or digitized information available in a sustainable manner is a major challenge. Concepts such as metadata, Meta Information, Linked Open Data (LOD), Big Data… are on the rise, but their meaning and content — let alone their implications in terms of social impact — are seldom questioned. Archival institutions thus face a myriad of challenges when sharing (meta)data with the scientific community and when planning to preserve information for future generations while maintaining data authenticity2 . ispartof: Archief- en Bibliotheekwezen in België vol:106 pages:5-13 status: published

Among the goals of the ARIADNE project (a digital infrastructure for archaeological documentation), two interrelated objectives are the design of new services (or the integration of existing ones) and the implementation of an intense training & hands-on program. Especially in the case of visual media (which are evolving at an impressing pace) services should be paired by user consultation and training, since a pre-requisite of any service activity is to consolidate knowledge on the potential of new media and the impact those could have in the selected discipline. The latter is the main objective of the TNA training program and of other user consultation events that have been organized in the first 18 months of the ARIADNE project. This short contribution presents the results of this activity and our plans for the design of new services.

Abstract Limited bandwidth is a strong constraint when efficient transmission of 3D data to Web clients and mobile applications is needed. In this paper we present a novel multi-resolution WebGL based rendering algorithm which combines progressive loading, view-dependent resolution and mesh compression, providing high frame rates and a decoding speed of million of triangles per second in JavaScript. The method is parallelizable and scalable to very large models. The algorithm is based on the local multi-resolution approaches provided by the community, but ad-hoc solutions had to be studied and implemented to provide adequate performances. In particular, a compression mechanism that reached very high compression rate without impact on rendering performance was implemented. Moreover, the data partition strategy was modified in order to be able to load different types of data (i.e. point clouds) and better adapt to the potentials and limitations of web-based rendering.