Reducing Optimism Bias in Incomplete Cooperative Games

In this paper, we investigate how to minimize the amount of information we learn about a cooperative game, while getting an accurate estimation of the Shapley value.

This paper was written by me, David Sychrovský, Jakub Černý and Martin Černý. It was accepted to the 23rd International Conference on Autonomous Agents and Multi-Agent Systems in Auckland, New Zealand.

The paper can be accessed in the Proceedings here. However, we recommend getting the version from arXiv, in which we fixed some minor mistakes:

1. The error bars were changed to be the standard error of the mean, as is more common in literature (previously was the standard deviation of the values).
2. A typo was fixed in the statement of Proposition 3.4 (an extra -v(N) after the second equals sign).