The problem in which multiple distrusting parties want to collaborate together and sample random variables which respect a certain joint distribution is called multi-party secure sampling. The talk will involve an introduction to the same; particularly the special case where parties do not have an offline setup. The characterization of distributions which can be sampled securely has been known for 3 parties for many years. This would be main content of the talk.

If time permits, we will also see a recent result of ours, a 4-party characterization for boolean distributions. This is a joint work with Keval Jain, Manoj Prabhakaran and Vinod Prabhakaran.