Recent news

First results from Corona Ad hoc survey: peoples' information sources in the crisis

CA-SG researcher Prof. Oliver Quiring and his team conducted an ad hoc survey on peoples’ information sources in times of the crisis. First results indicate that people rely on a mix of very diverse sources to obtain information on the pandemic: In the first days after the outbreak, people intensively used established news media, but also information shared by private contacts and official sources such as research institutions and public authorities. Moreover, the sources people rely on in the corona crisis significantly impacted peoples’ sense of community in the crisis, indicating that the way how information on the pandemic is presented considerably impacts the social cohesion in society.

New research paper: How to Best Predict the Daily Number of New Infections of COVID-19

Our team of CA-SG researchers (Lukas Jürgensmeier M.Sc., Kevin Stowe Ph.D., Prof. Dr. Bernd Skiera, and Prof. Dr. Iryna Gurevych) has just submitted our first research paper related to the Corona crisis. We compare the ability of several data sources, in particular Johns Hopkins University (JHU), Google search data and Twitter data, to predict the official number of new infections of Covid-19 and examine the need to complement the official numbers with additional predictions.

Our mission

Dealing with our complex world

With the increasing speed and ease of communicating multi-modal content, with the massive amount of digitized unstructured information, and with the technological and sociological changes, the world is more complex than ever before.

Our research focuses on analyzing and understanding unstructured data of

  • various origins (such as Tweets, news articles, web pages), and
  • various types (such as texts, images, videos, or spoken language).

Supporting time-critical political decisions

Especially in crises, decisions have to be made quickly, often under great uncertainty.

Often the problem is not the missing information, but the inability to process it and connect the dots. For example, images or videos could be temporally related to text messages or public information in order to identify critical situations early.

Our research focuses on supporting

  • rational and resource-rich content processing, and
  • policy makers to understand information in a context.

Building methods responsible at heart

Current methods for content analytics are prone to various flaws related to the values of our society. For example, they induce biases which lead to discrimination. They also lack an overall holistic approach to transparency and interpretability.

Our research in automatic content analytics thus pays special attention to

  • anti-discrimination,
  • privacy,
  • fairness, and
  • ethics.

Collaborating institutions