A5084395826- Semantic Web and Ontologies
- Natural Language Processing Techniques
- Digital Humanities and Scholarship
- ICT in Developing Communities
- Data Quality and Management
- Topic Modeling
- Biomedical Text Mining and Ontologies
- Music and Audio Processing
- Advanced Graph Neural Networks
- Web Data Mining and Analysis
- Library Science and Information Systems
- Service-Oriented Architecture and Web Services
- Scientific Computing and Data Management
- Digital and Traditional Archives Management
- Video Analysis and Summarization
- Mobile Crowdsensing and Crowdsourcing
- Advanced Database Systems and Queries
- Data Management and Algorithms
- Multimedia Communication and Technology
- Context-Aware Activity Recognition Systems
- E-Government and Public Services
- Geographic Information Systems Studies
- 3D Surveying and Cultural Heritage
- Innovative Approaches in Technology and Social Development
- Image Processing and 3D Reconstruction
Vrije Universiteit Amsterdam
2015-2024
Delft University of Technology
2023
University of Groningen
2023
University of Potsdam
2023
University of Stuttgart
2023
Goethe University Frankfurt
2023
Berlin Heart (Germany)
2023
TU Dortmund University
2023
Karlsruhe University of Education
2023
University of Amsterdam
2006-2021
In modern machine learning, raw data is the preferred input for our models.Where a decade ago scientists were still engineering features, manually picking out details we thought salient, they now prefer in their form.As long as can assume that all relevant and irrelevant information present data, design deep models build up intermediate representations to sift features.However, these are often domain specific tailored task at hand, therefore unsuited learning on heterogeneous knowledge: of...
In this document we describe the Amsterdam Museum Linked Open Data set. The dataset is a five-star representation and comprises entire collection of consisting more than 70,000 object descriptions. Furthermore,
Despite the pervasiveness of IoT domotic devices in home automation landscape, their potential is still quite under-exploited due to high heterogeneity and scarce expressivity most commonly adopted scenario programming paradigms. The aim this study show that Semantic Web technologies constitute a viable solution tackle not only interoperability issues, but also overall complexity modern scenarios. For purpose, we developed knowledge-based system which scenarios are result logical inferences...
Digital history is a branch of digital humanities concerned using ICT to improve study history. Linked Data provides way effective enriched access scientific texts about (historiographies). In this paper, we present method for connecting historiographical text the cloud. We and tools that use in each method's steps. focus on one extensive case study: an important work Dutch World War II historiography "Het Koninkrijk der Nederlanden de Tweede Wereldoorlog". describe digitization two sources...
End-to-end multimodal learning on knowledge graphs has been left largely unaddressed. Instead, most end-to-end models such as message passing networks learn solely from the relational information encoded in graphs' structure: raw values, or literals, are either omitted completely stripped their values and treated regular nodes. In case we lose potentially relevant which could have otherwise exploited by our methods. To avoid this, must treat literals non-literals separate cases. We also...
Knowledge graphs enable data scientists to learn end-to-end on heterogeneous knowledge. However, most models solely from the relational information encoded in graphs' structure: raw values, as literal nodes, are either omitted completely or treated regular nodes without consideration for their values. In case we lose potentially relevant which could have otherwise been exploited by our learning methods. We propose a multimodal message passing network not only learns structure of graphs, but...
Hybrid Intelligence (HI) is a rapidly growing field aiming at creating collaborative systems where humans and intelligent machines cooperate in mixed teams towards shared goals. A clear characterization of the tasks knowledge exchanged by agents HI applications still missing, hampering both standardization reuse when designing new systems. Knowledge Engineering (KE) methods have been used to solve such issue through formalization roles knowledge-intensive processes. We investigate whether KE...