Proceedings. Thirteenth Annual IEEE Conference on Computational Complexity (Formerly: Structure in Complexity Theory Conference
DOI: 10.1109/ccc.1998.694585
|View full text |Cite
|
Sign up to set email alerts
|

Nonrelativizing separations

Abstract: We show that MA EXP , the exponential time version of the Merlin-Arthur class, does not have polynomial size circuits. This significantly improves the previous known result due to Kannan since we furthermore show that our result does not relativize. This is the first separation result in complexity theory that does not relativize. As a corollary to our separation result we also obtain that PEXP, the exponential time version of PP is not in P=poly.

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1

Citation Types

0
64
0
1

Publication Types

Select...
4
3

Relationship

0
7

Authors

Journals

citations
Cited by 51 publications
(65 citation statements)
references
References 14 publications
0
64
0
1
Order By: Relevance
“…In particular, we do not know of any Karp-Lipton style collapse 5 with that setting of s. We may attempt to replace SAT with a seemingly harder language, say some language in a class C, where C is contained in the class for which we want to show a lower bound. Unfortunately this issue cannot be fixed this way; for otherwise we would not have the current gap between the known super-polynomial size lower bounds (which hold with respect to classes that are contained in the second level, namely MA-EXP (Buhrman, Fortnow, and Thierauf (1998))), and the known exponential-size lower bounds (which only hold with respect to classes that are in the third level of the hierarchy, namely E Σ P 2 (Kannan (1982))). Indeed, it was argued by Miltersen, Vinodchandran, and Watanabe (1999) that Karp-Lipton style collapses that are needed for Kannan's strategy hold with respect to size functions up to half-exponential (a function s is half-exponential if s(s(n)) ∈ 2 Θ(n) ) but do not seem to carry over to larger size bounds, to 2…”
Section: +ω(1)mentioning
confidence: 99%
See 1 more Smart Citation
“…In particular, we do not know of any Karp-Lipton style collapse 5 with that setting of s. We may attempt to replace SAT with a seemingly harder language, say some language in a class C, where C is contained in the class for which we want to show a lower bound. Unfortunately this issue cannot be fixed this way; for otherwise we would not have the current gap between the known super-polynomial size lower bounds (which hold with respect to classes that are contained in the second level, namely MA-EXP (Buhrman, Fortnow, and Thierauf (1998))), and the known exponential-size lower bounds (which only hold with respect to classes that are in the third level of the hierarchy, namely E Σ P 2 (Kannan (1982))). Indeed, it was argued by Miltersen, Vinodchandran, and Watanabe (1999) that Karp-Lipton style collapses that are needed for Kannan's strategy hold with respect to size functions up to half-exponential (a function s is half-exponential if s(s(n)) ∈ 2 Θ(n) ) but do not seem to carry over to larger size bounds, to 2…”
Section: +ω(1)mentioning
confidence: 99%
“…The first result of this flavor was given by Buhrman, Fortnow, and Thierauf (1998) who showed that if the class MA is equal to NP then NEXP ⊆ P/poly.…”
mentioning
confidence: 99%
“…The class of functions #EXP is defined analogously to #P, except with T a non-deterministic exponential-time machine. We will deal with a decision version of #EXP, PEXP, the set of problems solvable by nondeterministic Turing machine in exponential time, where the acceptance condition is that more than a half of computation paths accept [BFT98].…”
Section: Proof Notice That Ifmentioning
confidence: 99%
“…However, even non-uniform classes such as P/poly can be separated from large enough classes by means of diagonalization. Diagonalization combined with "arithmetization 1 " yields the best-known result along these lines: the result of Buhrman, Fortnow and Thierauf [14] that MA EXP is not contained in P/poly (and hence is also not contained in (non-uniform) TC 0 ). Is there hope that these techniques might lead to better lower bounds for TC 0 ?…”
Section: Tried and True Techniquesmentioning
confidence: 95%
“…Diagonalization is the canonical example of a "relativizing" proof technique, and even when when combined with arithmetization techniques the known separation results "algebrize" (using the terminology introduced by Aaronson and Wigderson [1]). Aaronson and Wigderson show that algebrizing proof techniques are not strong enough to prove that NEXP is not in P/poly (so that the lower bound of [14] mentioned above is close to the best that can be obtained using these techniques).…”
Section: Tried and True Techniquesmentioning
confidence: 99%