Learning systems in cultural heritage research
Call for Papers
Jaap Evert Abrahamse (1) | Erik Schmitz (2) | Rowin van Lanen (1, 3)
((1) Cultural Heritage Agency of the Netherlands | (2) Amsterdam City Archives | (3) Wageningen University and Research)
Keywords: pattern recognition, image recognition, HTR, big data analysis, machine learning, computer-aided research
Call: In the 1980s, the Austrian computer scientist Hans Moravec formulated his famous paradox, stating that reasoning, which is typical for humans, requires relatively little computation, as opposed to mobility and perception. Since then, computer power has grown rapidly. At the moment it is the concept of perception that proves an interesting though challenging feature for cultural heritage researchers. Present-day computers can generate more data, they can see sharper than humans and they can look through more data in less time. We see the emergence of more and more self-learning systems using human-computer interaction for text and pattern recognition and comparison in large digitized datasets like (handwritten) texts from archives, architectural designs, archaeological drawings and reconstructions, images of buildings, digital elevation maps or georeferenced historic maps.
In this session, we welcome researchers involved in the implementation of such or comparable techniques. Contributors are not limited to showing only their results, but are also encouraged to focus on work in progress and best practices. Furthermore, presenters are invited to reflect on the future development of these types of computer-aided research in heritage. In 2009 Moravec wrote that in the foreseeable future robot researchers will work alongside humans. Therefore, during the final discussion we would like to debate whether or not we expect computers to take the next step from recognition towards classification and interpretation. For instance, by connecting or linking information from datasets too large to humanly handle. These outcomes might fuel the more theoretical discussion: could computers be used not only to answer questions formulated in the past, but also help to define new questions and in doing so explore new territories?
Submission (open April 15, 2020)
Mind the guidelines