photo from http://www.huygensinstituut.knaw.nl
Our last lecture here in the Netherlands was given by Karina van Dalen-Oskam, head of the department of ICT Texts at the Huygens Instituut. She gave a fascinating introduction to digital and computational humanities and how she uses textual analysis to find and present new information about medieval texts. Another project — a very cool one — she has been involved in is the transcription, annotation and translation of Vincent van Gogh’s letters. The website, vangoghletters.org, is an amazing example of how digital resources can be used in literary studies.
Today we had another chance to marvel at the brilliance of Andrea Scharnhorst. Her lecture was about visual interfaces to knowledge, which she said have taken on more importance in information sciences. The reason, she said, is that with the flood of information available to the public now, visualizations can help people see patterns and important points more easily. The problem becomes finding the right visualization to present the information. Some good examples of information visualizations can be found on the Places & Spaces website.
Maarten Marx, an information systems professor at the University of Amsterdam, gave our guest lecture today about using technology to do automatic quantitative analysis of publicly available records. He talked mainly about his Political Mashup project, which aims to use data from newspapers and online sources to connect political topics and entities. As more and more text becomes available for analysis, having a way to automatically process and analyze it will be increasingly important. Text will be the main data of the future, Maarten said, and there will be new opportunities for research and quantitative analysis.
Another mind-blowing lecture by a super smartypants today — this time by Marieke van Erp. She talked about how tough, and how valuable, it is to get machines to learn how to use everyday speech. Really “smart” computers are still a long way off, but programming computers to be “lazy learners” (like humans — most of what we learn is not through intentional study) using techniques such as the k-nearest neighbor classifier. When computers are smart enough to infer meanings, they can take over some really tedious entry work, freeing researchers’ time and money for more important things. Marieke explained it much more elegantly, but that’s the general gist.
Before our guest lecture today, program director Trent Hill told us that last year’s group referred to our lecturer as “that really smart woman.” It’s an incredibly apt description. Andrea Scharnhorst gave a fascinating lecture about her research into article citations as a measure of journal quality. I’m sure it would have been even more fascinating if I had actually been able to understand more than half of what she was saying. It was not a language problem; her English was nearly perfect, as all of our lecturers’ English has been. Continue reading ‘Guest lecture: Andrea Scharnhorst’
Comments