Submodular Surrogates for Value of Information


  • Yuxin Chen ETH Zurich
  • Shervin Javdani Carnegie Mellon University
  • Amin Karbasi Yale University
  • J. Bagnell Carnegie Mellon University
  • Siddhartha Srinivasa Carnegie Mellon University
  • Andreas Krause ETH Zurich



Sequential Decision Making, Value of Information, Adaptive Submodularity, Decision Region Determination, Touch-based Localizatoin


How should we gather information to make effective decisions? A classical answer to this fundamental problem is given by the decision-theoretic value of information. Unfortunately, optimizing this objective is intractable, and myopic (greedy) approximations are known to perform poorly. In this paper, we introduce DiRECt, an efficient yet near-optimal algorithm for nonmyopically optimizing value of information. Crucially, DiRECt uses a novel surrogate objective that is: (1) aligned with the value of information problem (2) efficient to evaluate and (3) adaptive submodular. This latter property enables us to utilize an efficient greedy optimization while providing strong approximation guarantees. We demonstrate the utility of our approach on four diverse case-studies: touch-based robotic localization, comparison-based preference learning, wild-life conservation management, and preference elicitation in behavioral economics. In the first application, we demonstrate DiRECt in closed-loop on an actual robotic platform.




How to Cite

Chen, Y., Javdani, S., Karbasi, A., Bagnell, J., Srinivasa, S., & Krause, A. (2015). Submodular Surrogates for Value of Information. Proceedings of the AAAI Conference on Artificial Intelligence, 29(1).



AAAI Technical Track: Reasoning under Uncertainty