1: \begin{abstract}

2: This paper studies the problem of testing if an input $(\Gamma,\circ)$, where $\Gamma$ is a finite set of unknown size and $\circ$ 

3: is a binary operation over $\Gamma$ given as an oracle, is close to a specified class of groups.

4: Friedl et al.~[\emph{Efficient testing of groups}, STOC'05] have constructed an efficient tester using $\poly(\log|\Gamma|)$ queries for the class of 

5: abelian groups.

6: %, but left the case of subclasses of abelian groups open. In this paper we show that these latter problems are much harder:

7: We focus in this paper on subclasses of abelian groups, and show that these problems are much harder:

8: $\RomOmega(|\Gamma|^{1/6})$ queries are necessary to test if the input is close to a cyclic group, 

9: and $\RomOmega(|\Gamma|^{c})$ queries for some constant $c$

10: are necessary to test more generally if the input is close to an abelian group 

11: generated by $k$ elements, for any fixed integer $k\ge 1$. 

12: We also show that knowledge of the size of the ground set $\Gamma$ helps only for $k=1$, in which case we construct 

13: an efficient tester using $\poly(\log|\Gamma|)$ queries; 

14: for any other value $k\ge 2$ the query complexity remains $\RomOmega(|\Gamma|^{c})$.

15: All our upper and lower bounds hold for both the edit distance and the Hamming distance. These are, to the best of our knowledge,

16: the first nontrivial lower bounds for such group-theoretic problems in the property testing model and, in particular, they

17: imply the first exponential separations between the classical and quantum query complexities of testing closeness to 

18: classes of groups.\end{abstract}

19: