• DARIAH EU
• European Commission close
• ARIADNEplus close

L’articolo racconta l’esperienza del D4GNA - Dataset per il Geoportale Nazionale dell’Archeologia nato nell’ambito del progetto ARIADNEplus (Advanced Research Infrastructure for Archaeological Dataset Networking in Europe - plus), conclusosi lo scorso 31 dicembre. Il contributo parte dal contesto più ampio del progetto europeo per addentrarsi nel mondo dei dati archeologici italiani; la soluzione tecnologica, la standardizzazione, la dematerializzazione e la condivisione in rete dei dati sono i temi toccati in questo percorso che ci illustra il procedere verso un obiettivo virtuoso: il Geoportale Nazionale per l’Archeologia (GNA). Il GNA, realizzato dall’Istituto Centrale per l’Archeologia (ICA) e che sarà in rete dal 10 luglio 2023, è il punto di accesso nazionale per accogliere e consultare gli interventi archeologici svolti sotto la direzione scientifica del Ministero della Cultura (MiC), le indagini archeologiche condotte da università e altri enti di ricerca, nonché altre banche dati territoriali.

The ARIADNEplus project aggregated over 3.5 million metadata records describing archaeological sites. These records were converted and integrated to an RDF triple store using a data schema based on the CIDOC Conceptual Reference Model (CRM), then exported to an OpenSearch index for use with the search portal UI. Search is via 3 major facets - place, time and subject. This presentation primarily concerns the subject integration work which involved 59 local vocabularies containing over 19,000 subject terms in 16 different languages. A vocabulary matching tool allowed each data provider to define correctly formatted mappings from their own local vocabularies to a central common 'spine' vocabulary (Getty Art & Architecture Thesaurus, AAT). Using the resultant mappings, the metadata records were enriched with AAT identifiers. The AAT poly-hierarchical structure was incorporated into the search system, facilitating both multilingual cross search and hierarchical subject expansion. This approach led to a noticeable improvement in both recall and precision.

This deliverable presents the interim Dissemination and Communication Plan and Report 2 for the ARIADNEplus project. The report covers the second 18-month period (M18-M36, i.e. July 2020- December 2021) for the dissemination and communication activities undertaken by the lead team at PIN Scrl and all the (40+) partners who are involved in the project. During the first period, the project partners and their organisations were the main focus for dissemination and communication, with this policy continuing for the second period (M19-M36). Information and news have continued to be published on the website, which is aimed at a wider audience, with other methods of communication which have included partner’s mailing lists, social media channels and the Project Newsletter. There have not been any training events scheduled for period 2 and Transnational Access has remained suspended due to the current travel problems. Most of the activity by the project partners has been focussed upon the ingestion and publication of their datasets which started to appear in the public Portal from April onwards. It was agreed that the best method of dissemination would be national campaigns led by each of the partners who would target their local archaeological communities through their preferred communication channels when their datasets were published, backed up by the Project team via the website and Twitter. The ARIADNEplus website has met its M36 targets with Twitter and Facebook being major sources of referral along with partner websites and E-RIHS, the Research Infrastructure for Heritage Science (who has also been involved in the development of the Application Profile for this sub-domain). The updated Home page design was launched at the start of Period 2 with new content being added in the form of a Training Hub (providing training resources) and also a couple of webinars being made available. The Newsletter has a good opening rate at over 40%, and the subscriber list has grown slowly but steadily. Slideshare and Zenodo have also proven to be effective for promoting the project objectives with the three translations of the Guidelines to “Farify data management and make data reusable” into Czech, Portuguese and Turkish have generated over 1,000 views and 940 downloads. Twitter continues to be the most effective channel for dissemination. The Partners have participated in 20 conferences and events, giving presentations and taking part in round tables and workshops including CAA and EAA which were held virtually as were the great majority of events during this time. ARIADNEplus has also produced eleven articles and papers, the most recent drawing on the TNA training completed at the start of 2020 on CIDOC-CRM at PIN. The Communication and dissemination plan for M37-M48 involves an intensification of activities as well as specific targeting of the sub-domains as more data appears in the Portal, the pilots are published and training materials are prepared for a new section in the Training Hub. CAA and EAA are the two major conferences targeted by the project, along with DH2022 in Japan as this will promote ARIADNEplus in Asia.

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.

Im Allgemeinen haben Open Science-Praktiken nur langsam Auswirkungen auf die gangige archaologische Praxis. Eine Ausnahme bildet der offene Zugang zu Publikationen, der – zusammen mit offenen Daten und Methoden – zu den fur Archaolog*innen wahrscheinlich wichtigsten offenen Praktiken zahlt. Dieser Artikel stellt eine Auswahl von Archaologieprojekten vor, die offene wissenschaftliche Praktiken anwenden und fordern. Schlieslich wird es eine Diskussion uber einige der Fragen und Herausforderungen geben, denen sich die Disziplin auf ihrem Weg zu einer Open Archaeology gegenubersieht.

This book is a collection of seventeen papers which describe the impact that the ARIADNE project and its successor, ARIADNEplus (2019-2022) have had on the archaeological community, both in Europe and further afield. Each case study has been contributed by organisations involved in the ARIADNE Infrastructure who cover many countries from across Europe as well as Argentina and Japan. These papers were originally presented at the CAA Conference in Krakow, April 2019 and cover aspects such as data management, application of standards and guidelines, the use of CIDOC-CRM and Open Data to name but a few.

This book is a collection of seventeen papers which describe the impact that the ARIADNE project and its successor, ARIADNEplus (2019-2022) have had on the archaeological community, both in Europe and further afield. Each case study has been contributed by organisations involved in the ARIADNE Infrastructure who cover many countries from across Europe as well as Argentina and Japan. These papers were originally presented at the CAA Conference in Krakow, April 2019 and cover aspects such as data management, application of standards and guidelines, the use of CIDOC-CRM and Open Data to name but a few. Printed copies may be requested from Archaeolingua.

The Visual Media Service provides easy publication and presentation on the web of complex visual media assets (3D, RTI, High Resolution 2D Images). It is an automatic service that allows to upload visual media files on an server and to transform them into an efficient web format, making them ready for web-based visualization. The Visual Media Service is an end-to-end solution able to support the entire publishing process, from data uploading to presentation customization. The service provides a very easy interface (a simple web form) that allows the user to upload the visual media file; the data is then transformed in a web-compliant format, supporting multi-resolution encoding, compression, and progressive transmission. At the end of the data-processing phase the user receives an email containing the URL of the published asset. Through this URL the user can access a web-based presentation that enables content creators, but also end users, to interactively explore the published media. Finally, the Visual Media Service provides a full set of interactive tools for customizing the resulting web-based presentation.