Virtual Data Discussion w/ Hilda Rømer Christensen & Anna Rogers
Copenhagen Center for Social Data Science (SODAS), is pleased to announce that we are continuing with SODAS Data Discussions this fall.
SODAS aspirers to be a resource for all students and researchers at the Faculty of Social Sciences. We therefor invite researchers across the faculty to present ongoing research projects, project applications or just a loose idea that relates to the subject of social data science.
Every month two researchers will present their work. The rules are simple: short research presentations of ten minutes are followed by twenty minutes of debate. No papers will be circulated beforehand, and the presentations cannot be longer than five slides.
Associate Professor at Department of Sociology Hilda Rømer Christensen and Postdoc at SODAS Anna Rogers will present their work. Hilda Rømer Christensen works on the TInnGOproject analyzing Google image search results around transportation. Anna Rogers' main research area is Natural Language Processing. She works on interpretability and evaluation of deep learning models, as well as computational social science.
Hilda Rømer Christensen: Digital analysis as a research method for studies of transport, mobility and diversity?
The aim of this presentation is to present the TINNGO digital analysis as a possible research method for studies of transport, mobility and diversity. The aim has been to explore digital analysis as a new mode of evidence based knowledge production which enables comparative analysis of various gender diversity and transport discourses throughout Europe. More precisely the digital queries throw lights on affiliated networks related to terms such as smart mobility and passenger as well as well as gendered patterns of transport and employment. Moreover it is demonstrated that digital analysis is not an isolated tool, but is in need of interpretation and alignments with conceptual and contextualized analysis.
Anna Rogers: Can GPT-3 Use its Vast Knowledge?
The chief strength of the current deep-learning-based methods in Natural language processing is their ability to learn from large amounts of textual data; however, it is not clear whether this process can in principle result in human-like verbal reasoning ability. The latest development in scaling these methods is GPT-3, a language model with 175 billion parameters that was trained on 300 billion tokens of text. This work puts to test the reasoning abilities of GPT-3 in the "few-shot learning" mode: to what extent can it use the knowledge that it has seen in training?
The SODAS Data Discussion will take place at SODAS in Zoom from 11.00 am to 12.00 noon.
If you want to attend the event or just want to know more, please write Sophie Smitt Sindrup Grønning at firstname.lastname@example.org.