Interpretable Robust Decision Making

Interpretable decision making frameworks allow us to easily endow agents with specific goals, risk tolerances, and understanding. Existing decision making systems either forgo interpretability, or pay for it with severely reduced efficiency and large memory requirements. In this paper, we outline DeepID, a neural network approximation of Influence Diagrams, that avoids both pitfalls.

Authors

• Hal James Cooper Columbia University
• Garud Iyengar
• Ching-Yung Lin IBM Research

Keywords

• Interpretable agent modelling
• deep learning
• robustness