Note: An assertion concerning functionally complete algebras and NP-completeness
Theoretical Computer Science
An algebraic perspective on Boolean function learning
ALT'09 Proceedings of the 20th international conference on Algorithmic learning theory
ACM SIGACT News
Restricted two-variable FO + MOD sentences, circuits and communication complexity
ICALP'05 Proceedings of the 32nd international conference on Automata, Languages and Programming
An algebraic point of view on the crane beach property
CSL'06 Proceedings of the 20th international conference on Computer Science Logic
The multiparty communication complexity of exact-T: improved bounds and new problems
MFCS'06 Proceedings of the 31st international conference on Mathematical Foundations of Computer Science
Tight approximability results for the maximum solution equation problem over Zp
MFCS'05 Proceedings of the 30th international conference on Mathematical Foundations of Computer Science
The multiparty communication complexity of set disjointness
STOC '12 Proceedings of the forty-fourth annual ACM symposium on Theory of computing
Communication lower bounds using directional derivatives
Proceedings of the forty-fifth annual ACM symposium on Theory of computing
| Hi-index | 0.00 |
In this thesis, we address a number of issues pertaining to the computational power of monoids and semigroups as machines and to the computational complexity of problems whose difficulty is parametrized by an underlying semigroup or monoid and find that these two axes of research are deeply intertwined. We first consider the “program over monoid” model of D. Barrington and D. Thérien [BT88] and set out to answer two fundamental questions: which monoids are rich enough to recognize arbitrary languages via programs of arbitrary length, and which monoids are so weak that any program over them has an equivalent of polynomial length? We find evidence that the two notions are dual and in particular prove that every monoid in DS has exactly one of these two properties. We also prove that for certain “weak” varieties of monoids, programs can only recognize those languages with a “neutral letter” that can be recognized via morphisms over that variety. We then build an algebraic approach to communication complexity, a field which has been of great importance in the study of small complexity classes. We prove that every monoid has communication complexity O(1), &THgr;(log n) or &THgr;(n) in this model. We obtain similar classifications for the communication complexity of finite monoids in the probabilistic, simultaneous, probabilistic simultaneous and MOD p-counting variants of this two-party model and thus characterize the communication complexity (in a worst-case partition sense) of every regular language in these five models. Furthermore, we study the same questions in the Chandra-Furst-Lipton multiparty extension of the classical communication model and describe the variety of monoids which have bounded 3-party communication complexity and bounded k-party communication complexity for some k. We also show how these bounds can be used to establish computational limitations of programs over certain classes of monoids. Finally, we consider the computational complexity of testing if an equation or a system of equations over some fixed finite monoid (or semigroup) has a solution. (Abstract shortened by UMI.)