In 2009, using the
Fourier Checking problem, Aaronson claimed to construct the relativized worlds such that
BQP⊂BPPpath and
BQP⊂SZK. However, there are subtle errors in the original proof. In this paper, we point out the issues, and rescue these two separations by using more sophisticated constructions. Meanwhile, we take the opportunity to study the complexity classes
BPPpath and
SZK. We give general ways to construct functions which are hard for
SZK and
BPPpath (in the query complexity sense). Using these techniques, we give alternative construction for the oracle separation
BQP⊂SZK, using only Simon's problem. We also give new oracle separations for
PSZK from
BPPpath and
PSZK from
QSK. The latter result suggests that
PSZK might be strictly larger than
SZK.
In this work, we mostly concentrate on the hidden shift problem for non-injective functions. It is worthwhile to know that the query complexity of the non-injective hidden shift problem is exponential in the worst case by the well known bounds on the unstructured search problem. Hence, we can make this problem more tractable by imposing additional constraints on the problem. Perhaps the first constraint that comes to mind is to address the average case problem. In this work, we show that the average case non-injective hidden shift problem can be reduced to the injective hidden shift problem by giving one such reduction. The reduction is based on a tool we developed called injectivization. The result is strong in the sense that the underlying group can be any finite group and that the non-injective functions for which we have defined the hidden shift problem can have range in an arbitrary finite set. Using this tool, we simplify the main result of a recent paper by about the hidden shift problem for Boolean-valued functions by reducing that problem to Simon's problem. They also posed an open question which is subject to personal interpretation. We answer the seemingly most general interpretation of the question. However, we use our own techniques in doing so (the authors ask if their techniques can be used for addressing that problem). Another constraint that one can consider is to have a promise on the structure of the functions. In this work we consider the hidden shift problem for c-almost generalized bent functions. A class of functions which we defined that includes the generalized bent functions. Then we turn our attention toward the generalized hidden shift problem which is easier than injective hidden shift problem and hence more tractable. We state some of our observations about this problem. Finally we show that the average classical query complexity of the non-injective hidden shift problem over groups of form (Z/mZ)^n when m is a constant is exponential, which also immediately implies that the classical average query complexity of the non-injective hidden shift problem is exponential. We also show that the worst-case classical query complexity of the generalized injective hidden shift problem over the same group is high, which implies that the classical query complexity of the hidden shift problem is high.