Separations in Proof Complexity and TFNP

It is well-known that Resolution proofs can be efficiently simulated by Sherali--Adams~(SA) proofs. We show, however, that any such simulation needs to exploit huge coefficients: Resolution cannot be efficiently simulated by SA when the coefficients are written in unary. We also show that \emph{Reversible Resolution} (a variant of MaxSAT Resolution) cannot be efficiently simulated by Nullstellensatz (NS). These results can be interpreted in the language of total $\textsf{NP}$ search problems. We show that $\textsf{PPADS}$, $\textsf{PPAD}$, $\textsf{SOPL}$ are captured by unary-SA, unary-NS, and Reversible Resolution, respectively. Consequently, relative to an oracle, $\textsf{PLS}\not\subseteq\textsf{PPADS}$ and $\textsf{SOPL}\not\subseteq\textsf{PPA}$.