Editing Open Problems:66
Warning: You are not logged in. Your IP address will be publicly visible if you make any edits. If you log in or create an account, your edits will be attributed to your username, along with other benefits.
The edit can be undone.
Please check the comparison below to verify that this is what you want to do, and then save the changes below to finish undoing the edit.
Latest revision | Your text | ||
Line 1: | Line 1: | ||
{{Header | {{Header | ||
+ | |title=Distinguishing Distributions with Conditional Samples | ||
|source=bertinoro14 | |source=bertinoro14 | ||
|who=Eldar Fischer | |who=Eldar Fischer | ||
}} | }} | ||
− | Suppose we are given access to two distributions $P$ and $Q$ over $\{1,2, \ldots, n\}$ and wish to test if they are the same or are at least $\epsilon$ apart under the $\ell_1$ distance. Assume that we have access to ''conditional samples'': a query consists of a set $S \subseteq \{1,2, \ldots, n\}$ and the output is a sample drawn from the conditional distribution on $ | + | Suppose we are given access to two distributions $P$ and $Q$ over $\{1,2, \ldots, n\}$ and wish to test if they are the same or are at least $\epsilon$ apart under the $\ell_1$ distance. Assume that we have access to ''conditional samples'': a query consists of a set $S \subseteq \{1,2, \ldots, n\}$ and the output is a sample drawn from the conditional distribution on $A$. In other words, a conditional sample on $P$ given $S$ is drawn from the distribution where |
− | |||
− | |||
− | + | $$ \text{Pr}(j) = \begin{cases} \frac{p_j}{\sum_{i \in A} p_i} & j \in A \\ 0 & \text{otherwise} \end{cases} $$ | |
− | + | It is known that if one of the distributions is fixed, then a constant number ($f(1/\epsilon)$, where $f$ is an arbitrary function) of queries suffice to test the distributions. | |
− | |||
− | + | What can we say if both distributions are unknown? |