Massachusetts Institute of Technology
Computer Science and Artificial Intelligence Lab.
The Stata Center, Building 32
32 Vassar Street
Cambridge, MA 02139
United States of America
It is well known that the communication complexity of the "(monotone) Karchmer-Wigderson game" corresponding to a boolean function exactly captures the depth (and hence size) of the smallest (monotone) *formula* computing the function. The DAG model of communication complexity generalizes this connection and exactly captures the size of the smallest (monotone) *circuit* computing the function!
In this work we show that for any unsatisfiable CNF formula F that is hard to refute in the Resolution proof system, a gadget-composed version of F requires large DAG communication protocols. By the above connection, this implies that a monotone function associated with F has large monotone circuit complexity. [Our result also extends to monotone real circuits, which yields new lower bounds for the Cutting Planes proof system.] (joint work with Ankit Garg, Mika Göös, Dmitry Sokolov. [ECCC TR17-175]).