Jacco van Ossenbruggen appointed as Full Professor of Human-Centered Data Science
His research focuses on data-intensive applications where humans play a central role, e.g. where data used, directly concern people, where human judgement in the data analysis is of essential importance or because the result of the analysis needs correct interpretation by people. Van Ossenbruggen: “Our research continues to touch upon application areas wherever there’s too much data to manually process – where a problem or data are too complex or too subjective to be entirely left over to machines. In order to come up with good solutions, it is necessary to take the human centered role into consideration at the very beginning.”
An illustrative example is the recently founded Civic AI lab, where the VU, in collaboration with UvA, municipality Amsterdam and the Ministry of the Interior performs research into the use of artificial intelligence (AI) in the public administration. Many AI algorithms in the lab work surprisingly well for many entirely different sorts of data. Computer scientists may therefore feel easily tempted to abstract from the details to data as these “don’t matter anyway”. However, where data involves people or where results of such an algorithm lead to assumptions which can affect people, there can be talk of a wrong approach. Enthusiasm about an algorithm with a 96% correct score in the lab can be easily forgotten if practice later shows that 4% of errors systematically lead to e.g. more social inequality in a certain area. By properly understanding the problem and data all the way from the start, Civic AI aims to develop algorithms with a primary goal to decrease social inequality.
Another example is designing AI systems in the Cultural AI lab, where the VU cooperates with amongst others, Rijksmuseum, Koninklijke Bibliotheek (National Library of the Netherlands) and Beeld and Geluid (Sound and Vision). Through learning from data, AI systems unintentionally take over undesired patterns which later turn out to be hidden in the data, e.g. racial and gender bias. This can be approached in a different way: in the Netherlands a huge amount of heritage has been digitalized by museums, libraries and archives. Institutes and humanities researchers possess extensive knowledge about this heritage. They are often well aware of the bias processed in their data. The question is, how to use that knowledge to systematically train the algorithms in order for them to recognize the bias. Can AI systems learn that the stories about same objects or events can be told from a number of different perspectives? Van Ossenbruggen: “When once watching the 8 o’clock evening news, I saw how the visitors of the Exhibition on slavery left in tears – stories behind the exhibits made huge impact on them. However, these exhibits and stories are based on the same database where our research is performed on. This creates enormous expectations – yet, it is an immense challenge to investigate how we can do that in a responsible way.”
Jacco van Ossenbruggen defended his PhD in 2001 at the Computer Science department at the VU. Until recently he led a research group at the HCDA research group at CWI (National Research Institute for Mathematics and Computer Science) Amsterdam. Since 2018 he has led the User-centric Data Science research group at the VU. He is a co-founder of the Civic AI and Cultural AI ICAI labs. As an ODISSEI and CLARIAH management board member, Van Ossenbruggen has worked on improving the digital infrastructure for researchers in social science and humanities.
From September onwards Amsterdam Data Science will merge media channels with Amsterdam AI. Any online activities you are used to will continue on the Amsterdam AI channels. So please register below to stay up to date!
Amsterdam Data Science is excited to announce the next step in joining forces with Amsterdam AI. Together, we will support Amsterdam’s development as an international hub for Responsible AI.
The UvA Data Science Center is announcing a groundbreaking research program to align artificial intelligence (AI) for the interpretation of video data with human values and ethical principles.