On Boolean vs. Modular arithmetic for circuits and communication protocols
We compare two computational models that appeared in the literature in a Boolean setting and in an analog setting based on modular arithmetic. We prove that in both cases the arithmetic version can to some extend simulate the Boolean version. Although the models are very different, the proofs rely on the same idea based on the Schwartz-Zippel-Theorem.
In the first part we prove that depth d semi-unbounded Boolean circuits can be simulated by depth 2d + O(log d + log n) semi-unbounded arithmetic circuits, regardless of the size. This is an improvement on a similar construction in  that achieves depth 3d + O(log s + log n), where s is the size of the original circuit. Our construction is simpler and uses fewer random bits. In the second part we prove, that two-party parity communication protocols can approximate nondeterministic communication protocols. A strict simulation of one by the other is impossible as was shown in .
Unable to display preview. Download preview PDF.
- 1.Amos Beimel and Anna Gál. On arithmetic branching programs. Proc. Conf. on Computational Complexity, 1998 (Preliminary version: Technical Report 97-81, Center for Discrete Mathematics and Theoretical Computer Science 1997)Google Scholar
- 2.Carsten Damm, Matthias Krause, Christoph Meinel, and Stephan Waack. On Relations Between Counting Communication Complexity Classes. Journal on Computer System Sciences (to appear).Google Scholar
- 3.Anna Gál and Avi Wigderson. Boolean complexity classes vs. their arithmetic analogs. Random Structures and Algorithms. John Wiley & Sons, Inc., 1996. see also: Electronic Colloquium on Computational Complexity, Report 95-49, http://www.eccc.uni-trier.de/eccc/, 1995.Google Scholar
- 4.Avi Wigderson. NL/poly \(\subseteq\) ⊕L/poly. Proc. of the 9th Conference on Structure in Complexity Theory, pp. 59–62, 1994.Google Scholar
- 5.Rudolf Lidl and Harald Niederreiter. Introduction to finite fields and their applications. Cambridge University Press, 1986.Google Scholar
- 6.Rajeev Motwani and Prabhakar Raghavan. Randomized Algorithms. Cambridge University Press, 1995.Google Scholar
- 7.K. Mulmuley, U. Vazirani, and V. Vazirani. Matching is as easy as matrix inversion. In Proceedings of the 19th STOC, pages 345–354, 1987.Google Scholar
- 8.Klaus Reinhardt and Eric Allender. Making nondeterminism unambigous. Electronic Colloquium on Computational Complexity, Report 97-14, http://www.eccc.uni-trier.de/eccc/, 1997.Google Scholar