Social Foundations of Computation

Focus research areas include:

• Machine learning and algorithmic decision making in social contexts. Applying machine learning in social contexts runs into century-old unresolved questions about the limits of prediction and the predictability of social events. The need to optimize, learn, and intervene in dynamic environments challenges prevailing problem formulations and objectives in the field. 
• Models of social and economic interaction. Formal models of social action give a mathematical basis for reasoning about individuals, their actions and reactions, relationships, structures, and populations. Economic models come in when studying incentives, competition, and collaboration in applications of machine learning. 
• Normative goals and mitigation of harms. Histories of the politics and misuse of formal models inform proposed technical solutions. Discrimination and inequality in the allocation of opportunities and resources are major concerns as algorithms increasingly influence people’s life chances. A study of harms ties into a precise specification and pursuit of normative goals, such as fairness.
• Scientific foundations. The department confronts the ongoing methodological crisis in the empirical sciences with an emphasis on the collective practices of scientific communities including dataset creation, experimentation, and benchmarking. Reliability and validity, generalization, and extrapolation are fundamental subjects of research, studied in context rather than in statistical isolation. 

Situated within the Max Planck Institute for Intelligent Systems in Tübingen, the department extends the expertise of the institute in machine learning, artificial intelligence, robotics and physical systems in the direction of sociotechnical systems.  The department engages a growing community of scholars that study computation through a social lens, interrogating formal processes that sort, allocate, and classify. Integrating empirical and theoretical research, social foundations of computation also build on descriptive, normative, and critical perspectives from the humanities and social sciences.