• DARIAH EU
• 2021-2021close
• Open Access close
• ISTI Open Portal close

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.

Since the first ESFRI roadmap in 2006, multiple humanities Research Infrastructures (RIs) have been set up all over the European continent, supporting archaeologists (ARIADNE), linguists (CLARIN-ERIC), Holocaust researchers (EHRI), cultural heritage specialists (IPERION-CH) and others. These examples only scratch the surface of the breadth of research communities that have benefited from close cooperation in the European Research Area.While each field developed discipline-specific services over the years, common themes can also be distinguished. All humanities RIs address, in varying degrees, questions around research data management, the use of standards and the desired interoperability of data across disciplinary boundaries.This article sheds light on how cluster project PARTHENOS developed pooled services and shared solutions for its audience of humanities researchers, RI managers and policymakers. In a time where the convergence of existing infrastructure is becoming ever more important – with the construction of a European Open Science Cloud as an audacious, ultimate goal – we hope that our experiences inform future work and provide inspiration on how to exploit synergies in interdisciplinary, transnational, scientific cooperation. This article has been accepted for publication by EUP in the IJHAC: International Journal of Humanities and Arts Computing (https://www.euppublishing.com/loi/ijhac) International audience

Abstract In this article we present the first achievement of the Index Medii Aevi Geographiae Operum (IMAGO)—Italian National Research Project (2020–23), that is, the ontology we have created in order to formally represent the knowledge about the geographical works written in Middle Ages and Renaissance (6th–15th centuries). The IMAGO ontology is derived from a strict collaboration between the Institute of Information Science and Technologies (ISTI) of the Italian National Research Council (CNR) and the scholars who are involved in the project, who have supported ISTI-CNR in defining a conceptualization of the domain of knowledge. Following the re-use logic, we have selected as reference ontologies the International Committee on Documentation CRM vocabulary and its extension FRBRoo, including its in-progress reformulation, LRMoo. This research is included in a wider project context whose final aim is the creation of a knowledge base (KB) of Latin geographic literature of the Middle Ages and Renaissance Humanism in which the data are formally represented following the Linked Open Data paradigm and using the Semantic Web languages. At the end of the project, this KB will be accessed through a Web application that allows retrieving and consulting the collected data in a user-friendly way for scholars and general users, e.g. tables, maps, CSV files.