2024-03-29T13:42:27Z
https://zenodo.org/oai2d
oai:zenodo.org:4003499
2020-08-27T12:59:23Z
user-niaaproceedings1819
Babette Claassen
Jeroen Borst
Ingrid Vermeulen
Victor de Boer
Chris Dijkshoorn
2020-08-27
<p>Provenance research has become a key practice in the field of art<br>
market studies. The growing number of datasets and digital<br>
services around art-historical information presents new<br>
opportunities for conducting provenance research at scale. Usage<br>
of these new sources is hampered by the heterogeneity of<br>
information, worsened by temporal and cultural differences in<br>
documentation practices and its current digital storage/processing.<br>
In this paper we propose 1) a workflow model able to integrate<br>
provenance information from various sources, 2) a method to<br>
combine information from both on- and offline sources about art<br>
objects and auctions. We validate this method through a case<br>
study, where we investigate whether we can capture information<br>
from selected sources about an auction (1804), during which the<br>
paintings from the former collection of Pieter Cornelis van<br>
Leyden (1732-1788) were dispersed. The heterogeneous<br>
information acquired through the model might potentially be<br>
saved in a homogeneous database that can be processed to a<br>
Linked Open Data format. The idea behind this is that all the data<br>
gathered from both the online and offline sources will be<br>
processed in the same format and can help extend the information<br>
of available databases. Furthermore, by automating certain<br>
important steps in the process of provenance research, we are able<br>
to contribute to the facilitation and acceleration of this process<br>
altogether. The workflow model also provides a basic guideline<br>
for provenance research and future integration with with Linked<br>
Open Data process can lead to new potential for addressing<br>
relevant research questions for studies in the history of collecting<br>
and the art market<br>
</p>
https://doi.org/10.5281/zenodo.4003499
oai:zenodo.org:4003499
Zenodo
https://zenodo.org/communities/niaaproceedings1819
https://doi.org/10.5281/zenodo.4003498
info:eu-repo/semantics/openAccess
Creative Commons Attribution 4.0 International
https://creativecommons.org/licenses/by/4.0/legalcode
Cultural heritage
art markets
provenance
linked open data
knowledge acquisition
Linked Art Provenance
info:eu-repo/semantics/article
oai:zenodo.org:4003531
2020-08-27T12:59:23Z
user-niaaproceedings1819
Victor de Boer
Antske Fokkens
Christine Moser
Ivar Vermeulen
2020-08-27
<p>This volume contains the proceedings of the 2018/2019 edition of the Academy Assistant program<br>
organized by the Network Institute of Vrije Universiteit Amsterdam<br>
(http://networkinstitute.org/research/academy-assistants/). With this program the Network Institute<br>
aims to interest bright young master students for conducting scientific research and pursuing an<br>
academic career. The program brings together scientists from different disciplines; every project<br>
combines methods & themes from informatics, social sciences and/or humanities. For each project,<br>
two student research assistants with different research backgrounds work together. The projects<br>
result in papers and/or research proposals. The program started in 2010 with 4 projects funded by<br>
the KNAW. For the 2018/2019 edition, seven projects were selected to receive funding based on a review of<br>
submitted proposals by an independent committee. Each of these projects was then executed by two<br>
Academy Assistants during a 10-month period. The selected projects show a broad range of<br>
interdisciplinary research and a diverse range of topics<br>
</p>
https://doi.org/10.5281/zenodo.4003531
oai:zenodo.org:4003531
Zenodo
https://zenodo.org/communities/niaaproceedings1819
https://doi.org/10.5281/zenodo.4003530
info:eu-repo/semantics/openAccess
Creative Commons Attribution 4.0 International
https://creativecommons.org/licenses/by/4.0/legalcode
Frontmatter - Proceedings of the Netwerk Institute Academy Assistants programme 2018-2019
info:eu-repo/semantics/article
oai:zenodo.org:4003509
2020-08-27T12:59:23Z
user-niaaproceedings1819
Yanniek van der Schans
David Ruhe
Wido van Peursen
Sandjai Bhulai
2020-08-27
<p>This study examines linguistic variation within Biblical Hebrew<br>
by using Recurrent Neural Networks (RNNs) to detect differences<br>
and cluster the Old Testament books accordingly. Various linguistic<br>
features are analysed that are traditionally considered to be of importance in analysing linguistic variation. The traditional division<br>
of books as either Early Biblical Hebrew or Late Biblical Hebrew is<br>
hereby put to the test. Results show that RNNs are a fitting method<br>
for analysing the (morpho)syntax of a language. The model works<br>
well on both separate features, as well as all the features combined.<br>
On the basis of the results the RNNs provide, we propose that<br>
the diachronic approach to Biblical Hebrew is indeed plausible.<br>
The clusters generally hint to the scholarly division made in the<br>
diachronic approach to linguistic variation<br>
</p>
https://doi.org/10.5281/zenodo.4003509
oai:zenodo.org:4003509
Zenodo
https://zenodo.org/communities/niaaproceedings1819
https://doi.org/10.5281/zenodo.4003508
info:eu-repo/semantics/openAccess
Creative Commons Attribution 4.0 International
https://creativecommons.org/licenses/by/4.0/legalcode
Recurrent Neural Networks
Biblical Hebrew
Diachronic Liguistics
Computational Semantics
Clustering Biblical Texts Using Recurrent Neural Networks
info:eu-repo/semantics/article
oai:zenodo.org:4003493
2020-08-27T12:59:23Z
user-niaaproceedings1819
Ilze Amanda Auzina
Jenia Kim
Evert Haasdijk
Frank van Harmelen
Piek Vossen
2020-08-27
<p>Before a company enters a new business relationship it has to perform a background check, known as due diligence. It is commonly carried out by a human expert and involves screening a large amount of unstructured textual information (e.g. news articles), which is extremely labor intensive. We propose to automate this process, which would allow to, firstly, reduce the time needed for article screening, and, secondly, discover new insights about the network the company operates in. The solution includes (a) a classifier that detects articles containing negative events about the company of interest, and (b) a knowledge graph that combines the gained information with structured data sources. We report promising results of the novel approach to utilize semantic frames of the article’s predicates as features for the news article classification. Furthermore, we have successfully built a knowledge graph that combines information from different data sources. The proposed automated pipeline introduces a promising novel alternative for the commonly performed due diligence procedure.<br>
</p>
https://doi.org/10.5281/zenodo.4003493
oai:zenodo.org:4003493
Zenodo
https://zenodo.org/communities/niaaproceedings1819
https://doi.org/10.5281/zenodo.4003492
info:eu-repo/semantics/openAccess
Creative Commons Attribution 4.0 International
https://creativecommons.org/licenses/by/4.0/legalcode
Automated Due Diligence: Building Knowledge Graphs from News
info:eu-repo/semantics/article
oai:zenodo.org:4003501
2020-08-27T12:59:23Z
user-niaaproceedings1819
Lisa Vasileva
Edoardo Guerriero
Isa Maks
Kasper Welbers
2020-08-27
<p>Training models to perform stance identification requires large<br>
annotated corpora that are not always available, especially in a<br>
research setting where researchers must collect and annotate data<br>
from scratch. Active learning is a technique for reducing the number of texts that are required to train a model, but while it has been<br>
studied for decades, most of the papers that test the effectiveness<br>
of this approach in a natural language processing scenario rely<br>
on already made and well known datasets like the IMDB Movie<br>
Reviews Dataset. In this paper we test different active learning<br>
approaches in a scenario where task-specific training data needs<br>
to be developed from scratch. We collected and annotated tweets<br>
about vaccination to build our own dataset, and trained a model<br>
using different active learning pool-based strategies to see if active<br>
learning can be a viable strategy for building task specific stance<br>
identification models.<br>
</p>
https://doi.org/10.5281/zenodo.4003501
oai:zenodo.org:4003501
Zenodo
https://zenodo.org/communities/niaaproceedings1819
https://doi.org/10.5281/zenodo.4003500
info:eu-repo/semantics/openAccess
Creative Commons Attribution 4.0 International
https://creativecommons.org/licenses/by/4.0/legalcode
Building task-specific stance identification models: an evaluation of the active learning approach
info:eu-repo/semantics/article
oai:zenodo.org:4003542
2020-08-27T12:59:23Z
user-niaaproceedings1819
Victor de Boer
Antske Fokkens
Christine Moser
Ivar Vermeulen
2020-08-27
<p>This volume contains the proceedings of the 2018/2019 edition of the Academy Assistant program<br>
organized by the Network Institute of Vrije Universiteit Amsterdam<br>
(http://networkinstitute.org/research/academy-assistants/). With this program the Network Institute<br>
aims to interest bright young master students for conducting scientific research and pursuing an<br>
academic career. The program brings together scientists from different disciplines; every project<br>
combines methods & themes from informatics, social sciences and/or humanities. For each project,<br>
two student research assistants with different research backgrounds work together. The projects<br>
result in papers and/or research proposals. The program started in 2010 with 4 projects funded by<br>
the KNAW. For the 2018/2019 edition, seven projects were selected to receive funding based on a review of<br>
submitted proposals by an independent committee. Each of these projects was then executed by two<br>
Academy Assistants during a 10-month period. The selected projects show a broad range of<br>
interdisciplinary research and a diverse range of topics<br>
</p>
https://doi.org/10.5281/zenodo.4003542
oai:zenodo.org:4003542
Zenodo
https://zenodo.org/communities/niaaproceedings1819
https://doi.org/10.5281/zenodo.4003530
info:eu-repo/semantics/openAccess
Creative Commons Attribution 4.0 International
https://creativecommons.org/licenses/by/4.0/legalcode
Frontmatter - Proceedings of the Netwerk Institute Academy Assistants programme 2018-2019
info:eu-repo/semantics/article