In this thesis, we study small, yet important, circuit complexity classes within NC^1, such as ACC^0 and TC^0. We also investigate the power of a closelyrelated problem called Iterated Matrix Multiplication and its implications in low levels of algebraic complexity theory. More concretely, (1) We show that extremely modest-sounding lower bounds for certain problems can lead to non-trivial derandomization results. (a) If the word problem over S_5 requires constant-depth threshold circuits of size n^{1+epsilon} for some epsilon >0, then any language accepted by uniform polynomial-size probabilistic threshold circuits can be solved in subexponential time (and more strongly, can be accepted by a uniform family of deterministic constant-depth threshold circuits of subexponential size.) (b) If there are no constant-depth arithmetic circuits of size n^{1+\epsilon} for the problem of multiplying a sequenceof n 3-by-3 matrices, then for every constant d, black-box identity testing for depth-d arithmetic circuits with bounded individual degree can be performed in subexponential time (and even by a uniform family of deterministic constant-depth AC circuits of subexponential size). (2) ACC_m circuits are circuits consisting of unbounded fan-in AND, OR and MOD_m gates and unary NOT gates, where m is a fixed integer. We show that there exists a language in non-deterministic exponential time which can not be computed by any non-uniform family of ACC_m circuits of quasi-polynomial size and o(loglog n) depth, where $m$ is an arbitrarily chosen constant. (3) We show that there are families of polynomials having small depth-two arithmetic circuits that cannot be expressed by algebraic branching programs of width two. This clarifies the complexity of the problem of computing the product of a sequence of two-by-two matrices, which arises in several settings.
Recommendations
The Landscape of Communication Complexity Classes
We prove several results which, together with prior work, provide a nearly-complete picture of the relationships among classical communication complexity classes between $${\mathsf{P}}$$P and $${\mathsf{PSPACE}}$$PSPACE, short of proving lower bounds ...
Separating Complexity Classes Using Autoreducibility
A set is autoreducible if it can be reduced to itself by a Turing machine that does not ask its own input to the oracle. We use autoreducibility to separate the polynomial-time hierarchy from exponential space by showing that all Turing complete sets ...
Circuit Lower Bounds Collapse Relativized Complexity Classes
COCO '99: Proceedings of the Fourteenth Annual IEEE Conference on Computational ComplexitySince the publication of Furst, Saxe, and Sipser's seminal paper connecting AC0 with the polynomial hierarchy (Math. Systems Theory, 1984), it has been well known that circuit lower bounds allow you to construct oracles that separate complexity classes. ...