Problem 87: Equivalence Testing with Conditional Samples
Revision as of 22:57, 25 October 2017 by Ccanonne (talk | contribs) (Created page with "{{Header |source=focs17 |who=Gautam Kamath |title=Equivalence testing with conditional samples }} ''This is a continuation of this previous open problem...")
| Suggested by | Gautam Kamath |
|---|---|
| Source | FOCS 2017 |
| Short link | https://sublinear.info/87 |
This is a continuation of this previous open problem from the |Bertinoro workshop Bertinoro Workshop on Sublinear Algorithms.
For constant $\varepsilon>0$ it is known that the sample complexity of equivalence testing, in the conditional oracle model where the algorithm gets to condition the samples it receives on arbitrary subsets of the domain $[n]$, is $\Omega(\sqrt{\log\log n})$ and $O({\log\log n})$. [AcharyaCK-14,FalahatgarJOPS-15].
What is the exact dependence on the domain size of the sample complexity of equivalence testing in the conditional oracle model?
