主题: Structure Modeling for the Generation of Linguistic Insights
主讲人：Rebecca Anne Rrieda Kehlbeck
In recent years, topic modeling has become very important for document analysis of large data sets. The used algorithms, however, usually perform their calculations in a high-dimensional space, which makes user interaction with the model non-intuitive. Therefore, new kinds of interaction spaces are needed, that act as an abstraction layer between model and user, allowing them to incorporate the semantics of their domain knowledge into the model. Using semantically coherent structure modeling, we aim to improve the ability of users to analyze hypothesis they have about the corpus and gain insight into how context influences linguistic phenomena, for example questions.
Rebecca is a first year PhD student at University of Konstanz with a focus on information visualization. Within the research group "Questions at the interface" she works together with researchers from linguistics and computer science to visualize semantic structures of large text corpora.