Embedded Files
Motivated from the problems faced by under-served communities or in under-resourced settings, we aim to contribute to the theory of robust AI and optimization by developing tractable, near- optimal, and asymptotically exact approximation schemes applicable to broad problem classes.
How can we design models and algorithms for optimal decision-making that are scalable and robust to misspecifications, uncertainty, and unmodeled phenomena? How do we devise AI- and ML-based approaches that can leverage diverse and imperfect sources of data?
What data shall we collect and when to gain the most valuable information for downstream decision-making? What is the value of information? How can we address issues of shared responsibility between humans and algorithms (e.g., preference elicitation techniques)?
How to provide access to techniques of robust AI and robust optimization to practitioners? How to facilitate the deployment of these systems to address important societal challenges?
Report abuse