TY - GEN
T1 - Lifting to Bounded-Depth and Regular Resolutions over Parities via Games
AU - Alekseev, Yaroslav
AU - Itsykson, Dmitry
N1 - Publisher Copyright:
© 2025 Owner/Author.
PY - 2025/6/15
Y1 - 2025/6/15
N2 - Proving superpolynomial lower bounds on proof size in the proof system resolution over parities (Res(λ•)) remains a significant open challenge. A recent breakthrough by Efremenko, Garlik, and Itsykson (STOC 2024) established an exponential lower bound for regular Res(λ•). In this work, we introduce a lifting technique for regular Res(λ•), applicable to a wide range of formulas. Specifically, we develop a method that transforms any formula with large resolution depth into a formula requiring exponential-size regular Res(λ•) refutations. This transformation is achieved through a combination of mixing and constant-size lifting. Using this approach, we provide an alternative and improved separation between resolution and regular Res(λ•), originally proved by Bhattacharya, Chattopadhyay, and Dvorak (CCC 2024). We construct an n-variable formula with a polynomial-size resolution refutation of depth O(√n), yet requires regular Res(λ•) refutations of size 2ω(√n). Furthermore, we apply our technique to establish an exponential lower bound on the size of depth-cnloglogn Res(λ•) refutations, where n is the number of variables in the refuted formula, and c is a constant. The hard instances in this setting are Tseitin formulas lifted with the Maj5 gadget. Since even depth-n Res(λ•) captures all possible definitions of regular Res(λ•), our result yields an exponential lower bound for top-regular Res(λ•), resolving an open question posed by Gryaznov, Pudlák, and Talebanfard (CCC 2022).
AB - Proving superpolynomial lower bounds on proof size in the proof system resolution over parities (Res(λ•)) remains a significant open challenge. A recent breakthrough by Efremenko, Garlik, and Itsykson (STOC 2024) established an exponential lower bound for regular Res(λ•). In this work, we introduce a lifting technique for regular Res(λ•), applicable to a wide range of formulas. Specifically, we develop a method that transforms any formula with large resolution depth into a formula requiring exponential-size regular Res(λ•) refutations. This transformation is achieved through a combination of mixing and constant-size lifting. Using this approach, we provide an alternative and improved separation between resolution and regular Res(λ•), originally proved by Bhattacharya, Chattopadhyay, and Dvorak (CCC 2024). We construct an n-variable formula with a polynomial-size resolution refutation of depth O(√n), yet requires regular Res(λ•) refutations of size 2ω(√n). Furthermore, we apply our technique to establish an exponential lower bound on the size of depth-cnloglogn Res(λ•) refutations, where n is the number of variables in the refuted formula, and c is a constant. The hard instances in this setting are Tseitin formulas lifted with the Maj5 gadget. Since even depth-n Res(λ•) captures all possible definitions of regular Res(λ•), our result yields an exponential lower bound for top-regular Res(λ•), resolving an open question posed by Gryaznov, Pudlák, and Talebanfard (CCC 2022).
KW - Resolution over parities
KW - depth
KW - lifting
KW - lower bounds
KW - proof complexity
KW - regular resolution
KW - resolution width
UR - https://www.scopus.com/pages/publications/105009827758
U2 - 10.1145/3717823.3718150
DO - 10.1145/3717823.3718150
M3 - Conference contribution
AN - SCOPUS:105009827758
T3 - Proceedings of the Annual ACM Symposium on Theory of Computing
SP - 584
EP - 595
BT - STOC 2025 - Proceedings of the 57th Annual ACM Symposium on Theory of Computing
A2 - Koucky, Michal
A2 - Bansal, Nikhil
PB - Association for Computing Machinery
T2 - 57th Annual ACM Symposium on Theory of Computing, STOC 2025
Y2 - 23 June 2025 through 27 June 2025
ER -