Mehrere Institute der Österreichischen Akademie der Wissenschaften wollen die Lücke in der Ausbildung zum Thema maschinelles Lernen...
Prosopography is a very important research area in the humanities, as it enables researchers to answer a number of important questions about historical societies and the lives of people within them that are difficult to answer when only considering primary sources about individual persons or events alone. The various sources available to us about early modern European Catholic nuns and monks are among the most interesting for transformation into a modern prosopographical database. This is because monastic communities were highly interconnected throughout Europe, the nuns and monks came from a wide range of geographic and social backgrounds, and the data they produced about themselves and their peers was often of very high quality and well-structured.
For these reasons, various primary sources have already been transferred into the digital domain by different research efforts. NAMPI (“Nuns and Monks – Prosopographical Interfaces”) is a project that aims to alleviate some of the most severe drawbacks of the existing digital resources in the area of early modern religious prosopography. These are: a narrow focus and lack of metadata structures that makes content difficult to compare, and unclear technical standards that make digital data difficult to share and reuse for researchers.
NAMPI will create an RDF semantic database based on the factoid model ontology developed by Pasin & Bradley, which together with additional content-oriented ontologies allows an accurate portrayal of the lives of early modern nuns and monks. The data will be made available to researchers and for reuse by other projects through a modern and well-documented REST API and a website based on it. This will enable registered contributors to edit data and anonymous users to browse and query the prosopographical information and extract a complex and accurate picture of the persons represented in the database.
By the end of the project, NAMPI will incorporate data on around 82,000 persons from already available databases like ProDomo or Germania Sacra as well as from previously unpublished sources like the profession books of the Ursulines in Linz and Innsbruck. In this way NAMPI will contribute to making information on the lives of early modern nuns and monks available to researchers and the general public alike and help to further our knowledge about our past and our cultural heritage.
Constitutions contain the most important rules of our political systems and Constitutional Courts are the most powerful judicial institutions of European states. They are the guardians of the Constitutions and ensure, for instance, the respect for fundamental rights. Since the 20th century their importance is ever growing, which attracted a lot of interest from academia and beyond. Yet, despite numerous investigations, we know little about how different Constitutional Courts refer to each other. With the ECCN project, the researchers want to tackle this research gap.
It is now broadly understood that Constitutional Court decisions rarely form in a vacuum. Sometimes European Constitutional Courts rely on judgements from other European Constitutional Courts. Yet, more work is necessary to explore the extent of this phenomenon. In the ECCN project, the research team will measure empirically, if, and if yes how often, say, the Austrian Constitutional Court quotes or simply adopts the reasoning of other European Constitutional Courts such as for instance the German, the Polish or the Spanish Constitutional Court.
After exploring the direct and indirect quotations of European Constitutional Courts, the project team analyzes potential reasons for quoting other Constitutional Courts. The fact that European Constitutional Courts sometimes refer to other Constitutional Courts is interesting, because theoretically a Constitutional Court should refer in its decisions only to the Constitution and not to other Constitutional Courts. By analyzing the ECCN with computational methods, the project team wants to find reasons for this phenomenon.
Due to the big and multilingual corpus of the different European Constitutional Court decisions from 1952 to today, this task is challenging. For a single person it would be very difficult to compare all decisions of European Constitutional Courts, because of all the different European languages and the mere quantity of texts. Yet, the research team uses modern computational methods to carry out the analysis. More specifically, the team relies on their expertise in automated text analysis methods to tackle this task. Beyond the measurement of the citation network, that is the network of direct quotes between Constitutional Courts, machine learning allows them to generate a network of semantic text similarities. Both, citations and semantic text similarity, can be measured in texts with varying languages.
After digitalizing the ECCN, the following task is to explore conditions that explain patterns in this network. Hence, once the team has measured the network, they aim at explaining their results using inferential network analysis. With this approach, they do not only create original data, but also test established theories in a sophisticated and innovative way. In short, the project team digitalizes and explains the network of European Constitutional Court decisions.
The project “ONAMA” (“Ontology of Narratives of the Middle Ages”) deals with the creation of a computer-based system of relations between medieval narratological entities in texts and images.
The interdisciplinary ‘joint venture’ ONAMA is based on the extensive corpora of two long-term projects of the University of Salzburg, the Middle High German Conceptual Database (MHDBDB) and the image database REALonline of the Institute for Medieval and Early Modern Material Culture (IMAREAL). The two databases provide a sophisticated access to the medieval fine arts and literature: At REALonline a user is, among other things, able to choose a pictorial subject, to get access to a wide range of realisations of the motif appearing in works of art and on artefacts. The MHDBDB provides a comprehensive access to Middle High German literature and its vocabulary. Through its conceptual ontology it facilitates searches for motifs and themes and yields the respective text passages where they occur.
By connecting the data on the depicted elements and the thesauri of REALonline and the conceptual system of the MHDBDB, a language and media independent Ontology of Narratives of the Middle Ages will be formed. This innovative digital tool aims at overcoming the artificial disciplinary borderlines between text and image oriented studies and serves to answer many interdisciplinary and intermedial research questions. For example, ONAMA will examine how a narrative is realised through different objects and how the context of such “materialisations” influences the narratological structures. Thus it will reveal the impact a change of medium or the information carrier has on the narrative.
Users can simultaneously access detailed annotations on narratives in images and texts via the ONAMA frontend. This facilitates the identification of references or differences within the widely diversified corpora to medieval sources in the two databases. The narrative patterns in texts and images are annotated and visualized according to contemporary digital standards, which will provide a more solid empirical evaluation of the material. All data generated within ONAMA will be made available to the scholarly and scientific community under the Creative Commons CC BY-NC-SA license and can be used for example as a basis for queries about narratives in other digital corpora.
The project "Text Mining Medieval Mining Texts" (T.M.M.M.T) deals with two medieval mining sources which are of central importance for medieval mining in the alpine regions of Schwaz and Rattenberg in Tyrol (Austria). One of the two sources is unedited while the other is available in a publication.
The main objective of the research project is the extraction of information related to the legal relationships between people, claims, mines and the localization of the mines in the landscape. The central link to represent the relationships between people, pits and localization are names: names of places, people, tunnels, shafts and mining areas within the historical sources allow the representation of mutual relationships.
The extracted information will be represented in a semantic machine readable format (RDF) that allows the integration of other sources related to the subject and the visualisation of the spatial and temporal development of a mining district almost 600 years ago.
A major task is the creation of a searchable, further processable version of the medieval and handwritten documents "Verleihbuch der Rattenberger Bergrichter" (Hs. 37) and “Schwazer Berglehenbuch” (Cod. 1587). Both historical sources are currently stored by the Tyrolean Regional Archives (TLA) in Innsbruck, Austria. The handwritten medieval document Hs. 37 comprises precisely 347 pages and was written between 1460 and 1463. It contains the official claims to mine ores and minerals in the mining district of Brixlegg-Rattenberg (Tyrol, Austria). It also contains information about the timber industry and its connection to mining. The historical document “Schwazer Berglehenbuch” (Cod. 1587) comprises 408 pages was partly transcribed and published as scholarly edition by Wolfgang Tschan in 2009. Its content is similar to the content of Hs. 37 but it focuses on the mining district Falkenstein in Schwaz (Tyrol, Austria).
The process of linguistic annotation also includes the development of a suitable methodology for Named Entity Recognition and Named Entity Linking. Therefore, it is necessary to create gazetteers containing historical German orthography for place names and personal names. This data is available from the principal investigator’s dissertation as well as from other already transcribed historical sources (Hs. 12) stored by TLA. Another source for personal names will be the dissertation of Yvonne Kathrein (2011) and its database. Other sources will be historical maps and high resolution digital elevation models.
1. Digitized version of the medieval sources
2. Gazetteer of historical mining places
3. Methodology for Named Entity Recognition/Linking within historical documents
4. Semantic representation in RDF
5. Visualization of the temporal development of a mining district
6. Open access/online dissemination of the results 1-5
The Viennese avant-gardes of the 1950s and 1960s enjoy a superlative international reputation. Works from Viennese Actionism, the Vienna Group, or the Expanded Cinema are represented in the world's most important collections of contemporary art. As yet, however, Austria has not had an institutional hub that corresponds to this resonance. The Atlas of the Viennese Avantgardes (AVA) will close this gap and establish an internationally networked digital centre at the Austrian Academy of Sciences for researching and disseminating knowledge about the Viennese avant-gardes.
A central feature of the online platform will be filmed interviews with the still-living proponents of the postwar avant-gardes. The interviewees range from Valie Export and Peter Weibel, to Friedrich Achleitner and Oswald Wiener, to Friedrich Cerha and Wolf D. Prix. Thus, the project will secure invaluable, irreplaceable stocks of knowledge and Austrian cultural heritage.
The Atlas of the Viennese Avantgardes (AVA) will offer an interested public as well as the sciences an orientation, and it will create a space for the presentation of applicable research. On the one hand, it will focus on the Viennese avant-gardes as the heirs of Viennese Modernism around 1900 (Klimt, Schiele, Schönberg, Wittgenstein, etc.). On the other hand, research and dissemination will be directed towards the question of what connections to the present and social benefits can be derived from the works and thoughts of the avant-gardes.
How does a Twitter tweet, a Facebook post or a YouTube video shape how we feel about a public park? And what if we consume such media content in real-time and on-site using augmented reality technologies? How would this augmented reality change the way we use public places, feel about security, interact with others and how would it ultimately affect the cohesion of society? And if we react highly emotional to digital media portraying public places positively or negatively, how susceptible are we then to manipulation?
We are living in the midst of a spatial revolution where online and offline worlds increasingly converge. Digital media already shape our perception of spaces, e.g., through navigation systems that structure our geographic reality, through social media that bombard us with geo-tagged photos, videos and comments of cities, restaurants, or attractions. Yet, in face of the recent leap in augmented reality technology, our perception of space will once again profoundly change. The research gap on how this development affects individuals and society at large is still wide. Meanwhile, the field is lacking methods to adequately measure and understand how digital content impacts people’s affective-emotional experience of spaces. The project DigitAS – acronym for “The Digital, Affects and Space” – will develop new methods to tackle these pressing issues and will establish the infrastructure for further collaborative research and education in the field.
By selecting two highly debated public places in Austria as test sites – Rapoldi Park in Innsbruck and Venediger-Au-Park in Vienna – DigitAS will contribute a much-needed academic perspective to current debates on security and policing of public spaces. Results of the project will be shared with the public and authorities in two scenario workshops, in Innsbruck and Vienna, respectively. As part of the workshops, the DigitAS team seeks to discuss desirable future scenarios of development and necessary regulation of augmented and virtual reality with all invited stakeholders.
Network-based approaches play an increasingly important role in the analysis of data. Especially in the Digital Humanities (DH), network models have gained importance in recent years because more and more data-based and data-driven research is carried out and the amount of data is increasing (e.g. Big Data).
The project bridges the fields of linguistics, digital humanities and computer science in order to explore the diachronic dynamics of lexical networks on the basis of large-scale authentic language data. The project will reuse language data that is already available at the ACDH-OeAW, namely the Austrian Media Corpus (AMC) and the Corpus of Austrian Parliamentary Records (ParlAT). The AMC covers the entire Austrian media landscape of the past 20 years and contains 40 million texts (more than 10 billion tokens). The ParlAT corpus covers the Austrian parliamentary records of the last 20 years with more than 75 million tokens.
From the linguistic perspective, the project will explore the diachronic dynamics of lexical networks and discuss networks-based methods for diachronic linguistics. From the point of view of computer science, the project will apply network theory to a big amount of diachronic linguistic data and discuss new methods for the automatic analysis and comparison of these networks. Furthermore, the project will enrich the already available digital toolbox with a freely available tool for network analysis and visualisation and will enhance already existing data with additional annotations.
Economy and society are undergoing transformations that are increasingly shaped by innovations in the areas of robotics, algorithms and artificial intelligence. Automated decision systems perform the transactions in high frequency trading. The algorithms of search engines and social networks automatically process information dissemination on the Internet. Self-driving cars, self-directed energy supply in smart cities, and self-organized production networks for advanced manufacturing point to varieties and emerging fields of application for automation. Automation provides a broad range of potential benefits for social innovation and macro-economic growth. But at the same time automation is accompanied by several risks such as the loss of jobs, manipulation, security threats and privacy violations.
The emergence of risks of automation inevitably raises questions regarding responsibility. Who is liable if self-driving cars hit pedestrians? Who can be hold to account if chat bots on the Internet make racist statements? Who must respond to a critical public if automated decision systems contribute to discrimination? Numerous important questions regarding responsibility are not are not resolved yet. Responsibility gaps are frequently mentioned and this carries the danger that responsibility is denied or passed on, to users, technologies, or other actors, such as competitors for instance.
In this period of change and uncertainty, media reporting on risks and responsibilities plays an important role in shaping public perception and opinion formation on automation. The attribution of responsibility is of central relevance for the construction of responsibility. The proposed project will therefore investigate media coverage of risks and responsibilities on automation. The analyses will investigate which fields of automation (e.g. autonomous driving, curing robots, internet algorithms, etc.) and which risks of automation (jobs, security, surveillance, etc.) are subject to media reporting. It will also investigate which actors (industry, state, users, technologies, etc.) are held responsible for damages and for the development of solutions.
The project develops an innovative approach based on a responsibility concept from the philosophy of technology combined with (semi-)automated empirical content analyses of media reporting on automation. Hence, the project answers traditional responsibility questions in the humanities and social sciences by means of new and innovative methods from the digital humanities. The project is carried out by a interdisciplinary consortium with expertise in philosophy of technology, media- and communication science, computer science and advanced data processing. With regard to questions at the intersection of automation and responsibility this collaboration is unique in Austria.
Several hundreds of early medieval (app. 600-1100 AD) burial grounds and graveyards are known from the area of present day Austria and were analysed archaeologically and/or anthropologically. The earliest ones were dug out more than 150 years ago. Since then, methods and documentation of archaeological and anthropological research, as well as new analyses and ways of presenting the results, were developed. But up until today the most common way of presenting these data are printout publications. Only rarely texts and tables are published as Open Access Document, e.g. in form of PDFs. But structured data, enabling further research, are not available to the public nor the research community. The THANADOS-project aims at developing an online repository of the already published early medieval burial grounds in Austria. In it, three different disciplines - archaeology, anthropology and digital humanities - will be combined. The graves known and published to date will be inserted into the Open Source Database OpenAtlas, which models the data according to the international well-established CIDOC CRM standard. For that purpose THANADOS accesses already digitised data from bygone projects carried out by the project members, that will be published as Open Data for the first time. For that, a user interface with an appropriate entry workflow will be developed. With this, grave finds from respective locations can be entered in the database, connected to information about the human remains as well as grave goods and/or other related finds in a standardised way. Next to textual summaries and categorial data, pictures (photos, drawings) can be stored as well as georeferenced GIS-geometries. An interactive web-application for presenting the data will be developed by the team members of THANADOS in close cooperation with specialised colleagues of the Austrian Centre for Digital Humanities (ACDH) of the ÖAW. This web-application, attainable with every modern web browser, will allow the users to explore the data, search in the data pool and download raw data as well as specific search query reports in established formats. The cartographic and data-centered user interface complement each other smoothly. From a technical point of view, THANADOS is completely based on Open Source technology. The developments will be included in the OpenAtlas project under Open Source licenses. Depending on their categories, all data will be available for the public under Creative Commons licenses or as linked open data.
The project aims at building an open online repository of archaeological and anthropological data of early medieval burial grounds in nowadays Austria. Furthermore, this can be used as a model for other regions, epochs and topics.
No activity in healthcare is possible without language: All of the various settings in which healthcare is done rely on different forms of communication. Previous studies suggest that implicit forms of discrimination in healthcare communication persist and that patients are often discriminated depending on different social factors such as age, status, nationality or gender. This mostly occurs on an implicit, unconscious level; thus, discrimination is difficult to grasp. The project MedCorpInn wants to look for patterns of language use which are connected to such biases by investigating a large data set. The pre-existing digital data collection (corpus KARBUN) consists of 100.000 pre-anonymized radiology reports from the University Clinic of Innsbruck. Since there are only few similar corpora which consist of clinical texts, this data set can be referred to as unique.
The project aims to technically improve and develop the existing data set and to analyse the corpus with different tools and methods. For the technical improvements, the project group will work closely with their partners from DBIS (Databases and Informations Systems, UIBK). This collaboration will include the enhancement of the extensive metadata (age, gender, type of insurance, mode of examination etc.) as well as measures for automated data processing and for further anonymization. Moreover, the existing part-of-speech annotation will be improved; in this way e.g. the most frequent parts of speech can be found within the corpus.
Different corpus linguistic and discourse analytic methods and tools can be applied with the ready-made data. Thus, the applicants want to find out if differences of language use (if any) are connected to social factors. Therefore, the texts will be separated along the lines of certain categories (e.g. female/male, private insurance/public insurance etc.) and investigated in regards to statistically significant linguistic differences and patterns.
Furthermore, gender medical questions can be examined within the corpus. E.g. if medical procedures are linked with certain social factors, or if there are gender specific differences with regards to the precision of the measurements (e.g. of lengths/diameters of organs, tumours or injuries) and if yes, why.
MedCorpInn wants to find new methods of detecting structural biases on the linguistic surface of large datasets. Also, the project aims to contribute to the medical practice by creating ideas for guidelines to eliminate implicit biases and discriminations which are potentially harmful to patients.
The enlightened Spectator press of the 18th century constitutes an important cultural heritage of the world. It complied with the democratic ideal of disseminating cultural and moral issues, techniques and practices within a non-academic audience, popularizing enlightened ideas, such as cosmopolitanism, tolerance, intellectual criticism, self-reflection and social responsibility. Based on the existing text corpus of the digital scholarly edition of the Spectators (http://gams.uni-graz.at/mws), this cooperation between the Institute for Interactive Systems and Data Science, Graz University of Technology, the Know-Center GmbH Graz, and the Centre for Information Modelling - Austrian Centre for Digital Humanities (ZIM-ACDH) and the Institute for Romance Studies, both at the University of Graz, aims at investigating this multilingual corpus with computer-aided methods of quantitative text-analysis.
Quantitative methods of data analysis are becoming increasingly relevant in Digital Humanities due to the ever-increasing amount of data. Hence, data analysis is also becoming part of the methodological discourse (not only) in literary studies. What Franco Moretti designates as distant reading and Matthew Jockers subsumes under the term macroanalysis is aimed at the automated grouping and classification of large amounts of data without a previous examination using the close reading method.
The main objective of DiSpecs is to investigate how quantitative text-analysis methods are useful and efficient for the analysis of a multilingual text corpus. Computational methods will be developed to enhance and improve results gained from existing close reading approaches, with the goal of answering literary and linguistic research questions in a partly automated and also generally applicable and hence transferable way.
In the course of DiSpecs, it is expected that revelations about the shift of topics over time and geographic distance as well as stylistic features will allow for the formulation of statements about prominent trends in the 18th century journals. In particular, the quantitative analysis of the Spectator discourses aims to enhance and improve the studies on micro-narration regarding the repetition of motifs throughout different journals. Another focus is on the transnational-transfer and development of this literary genre, while keeping geographical, cultural and temporal specifics under constant consideration.
The results of the methodical studies will be incorporated into the existing digital edition of the Spectators in the form of visualizations and will also be used to improve existing search strategies.These additions will serve to promote the use of the digital edition not only as a valuable resource, but also as an actual research tool for literary scholars.
This project focuses on the development of the city of Graz in the "long 19th century" (1789-1914). A geographic information system will be established in order to analyze and visualize the transformation of the city through time as well as its perception thereof during the process of urbanization. Therefore, it combines maps with written sources (travel guides and reports), illustrative sources (postcards, images, graphic collection, early photography) as well as the latest historical research data, using existing digital assets, which will be complemented by not yet digitized sources. Thereby, BeUrb lays a solid foundation for the analysis of urban development in a period that is characterized by its rapid and radical changes.
Furthermore, the GIS provides a tool to identify places and structures, which had a strong impact on the city's evolution until today. Hence, the project not only gives an insight into the historical development but also provides approaches for a better understanding of the present-day city.
Apart from the scholarly use of such a system, the project will implement a web-platform which will also be open for private or touristic use, that can be accessed with mobile devices on site and will be implemented in future exhibitions of the City of Graz Museum. Thus, as a collaboration of the University of Graz, GrazMuseum and City Archive Graz, the project output will benefit scholarly research that reaches into society, but also further community engagement and the knowledge transfer to local schools and international tourists.
Recent years have seen a rise in new digital technologies and tools, and their combined use to access and analyse data in established but also in new and innovative ways. Increasingly, drawing knowledge from data collections or archives is not just at the centre of scientific work anymore, but has also seen other actor groups from society, industry or policy making become engaged in the process. ChIA – Accessing and Analysing Cultural Images with New Technologies is a collaborative effort between the Austrian Centre for Digital Humanities (ACDH-OeAW) and Europeana Local - Österreich (Europeana Local – AT) that aims to engage and test new technologies against the background of a selected dataset of food images for the benefit of accessing and analysing cultural data.
The ChIA project is innovative in that it proposes to use combinations of novel computational methods (visual search, chatbot technologies) and state-of-the-art tools (semantic thesauri) that have previously not been used together for cultural image access and analysis in the digital cultural heritage collection of Europeana Local - AT. With 2018 being the European Year of Cultural Heritage, science and technology for the benefit of cultural heritage plays a major role in both physical and digital realms. Besides, the topic of food is certainly a fundamental part of mankind’s cultural wealth and has a firm place in UNESCO’s Tangible and Intangible Cultural Heritage.
While the project aims to improve Europeana’s data access by creating an intermediate layer architecture to integrate and apply the proposed digital tools, it also enables increased and enriched knowledge insights for both scientists and citizens. Additionally, the project functions as a test-bed and playground for experimenting with Artificial Intelligence (AI) in a digital cultural context, where experimental settings and knowledge design practices open perspectives to other connected disciplines and serve as examples for similar future undertakings that aim to apply and enhance digital methods within the framework of cooperation projects.
ChIA is realised within the Open Innovation Research Infrastructure (OI-RI) at exploration space @ ACDH-OeAW, following principles of the Open Innovation Strategy for Austria, and further capitalises on knowledge by connecting with larger existing data infrastructures such as DARIAH, EGI and E-RIHS.