The complexity of joint computation
Scott Aaronson, Andrew Drucker
Abstract
Joint computation is the ubiquitous scenario in which a computer is presented with not one, but many computational tasks to perform.A fundamental question arises: when can we cleverly combine computations, to perform them with greater efficiency or reliability than by tackling them separately?This thesis investigates the power and, especially, the limits of efficient joint computation, in several computational models: query algorithms, circuits, and Turing machines.We significantly improve and extend past results on limits to efficient joint computation for multiple independent tasks; identify barriers to progress towards better circuit lower bounds for multiple-output operators; and begin an original line of inquiry into the complexity of joint computation.In more detail, we make contributions in the following areas:Improved direct product theorems for randomized query complexity: The "direct product problem" seeks to understand how the difficulty of computing a function on each of k independent inputs scales with k.We prove the following direct product theorem (DPT) for query complexity: if every T-query algorithm has success probability at most 1-e in computing the Boolean function f on input distribution p, then for a < 1, every aeTk-query algorithm has success probability at most (2 6(1 -E)) in computing the k-fold direct product f k correctly on k independent inputs from p.In light of examples due to Shaltiel, this statement gives an essentially optimal tradeoff between the query bound and the error probability.Using this DPT, we show that for an absolute constant a > 0, the worst-case success probability of any aR 2 (f)k-query randomized algorithm for f k falls exponentially with k.The best previous statement of this type, due to Klauck, Spalek, and de Wolf, required a query bound of O(bs(f)k).Our proof technique involves defining and analyzing a collection of martingales associated with an algorithm attempting to solve f*k.Our method is quite general This thesis was a long time in the making, and accordingly is indebted to many people.First of all, I would like to thank my advisor, Scott Aaronson, for the innumerable forms of support he has shown me over the years it's been my privilege to know him.Scott has always treated me as a colleague and an equal.He has always enthusiastically invited me to explore the many wonderful ideas that have excited him over the years, and has taught me a huge amount about computational complexity and allied fields.At the same time he has placed faith in me to plot my own course of research.My collaborations with Ronald de Wolf were in many ways akin to a second advisorship.In addition to sharing his scientific expertise, he generously shared his time to help me develop as a writer and editor.His formidable skills and work ethic have served as a model for me ever since.I am grateful to Russell Impagliazzo for his kind support and guidance during my first year of graduate study at UC San Diego.I thank Michael Sipser and Scott Aaronson for each giving me valuable teaching opportunities at MIT, and for leading by example with their own inspiring teaching.Fellow students have supported me and taught me a great deal over the years.Their guidance has come in many forms-from mathematical discussions, to practical advice about school and life.(I regard many of these talented individuals more as role models than as peers.)I have also valued their friendship and camaraderie through the academic process.A very partial list of fellow students who've given me valuable support and advice includes: