On Protocols for Monotone Feasible Interpolation

Author:

Folwarczný Lukáš1ORCID

Affiliation:

1. Institute of Mathematics, Czech Academy of Sciences and Computer Science Institute, Charles University

Abstract

Dag-like communication protocols, a generalization of the classical tree-like communication protocols, are useful objects in the realm of proof complexity (most importantly for monotone feasible interpolation) and circuit complexity. We consider three kinds of protocols in this article ( d is the degree of a protocol): IEQ- d -dags: feasible sets of these protocols are described by inequality which means that the feasible sets are combinatorial triangles; these protocols are also called triangle-dags in the literature, EQ- d -dags: feasible sets are described by equality, and c -IEQ- d -dags: feasible sets are described by a conjunction of c inequalities. Garg, Göös, Kamath, and Sokolov (Theory of Computing, 2020) mentioned all these protocols, and they noted that EQ- d -dags are a special case of c -IEQ- d -dags. The exact relationship between these types of protocols is unclear. As our main contribution, we prove the following statement: EQ-2-dags can efficiently simulate c -IEQ- d -dags when c and d are constants. This implies that EQ-2-dags are at least as strong as IEQ- d -dags and that EQ-2-dags have the same strength as c -IEQ- d -dags for c ≥ 2 (because 2-IEQ-2-dags can trivially simulate EQ-2-dags). Hrubeš and Pudlák (Information Processing Letters, 2018) proved that IEQ- d -dags over the monotone Karchmer-Wigderson relation are equivalent to monotone real circuits which implies that we have exponential lower bounds for these protocols. Lower bounds for EQ-2-dags would directly imply lower bounds for the proof system R(LIN).

Funder

GAUK

Publisher

Association for Computing Machinery (ACM)

Subject

Computational Theory and Mathematics,Theoretical Computer Science

同舟云学术

1.学者识别学者识别

2.学术分析学术分析

3.人才评估人才评估

"同舟云学术"是以全球学者为主线,采集、加工和组织学术论文而形成的新型学术文献查询和分析系统,可以对全球学者进行文献检索和人才价值评估。用户可以通过关注某些学科领域的顶尖人物而持续追踪该领域的学科进展和研究前沿。经过近期的数据扩容,当前同舟云学术共收录了国内外主流学术期刊6万余种,收集的期刊论文及会议论文总量共计约1.5亿篇,并以每天添加12000余篇中外论文的速度递增。我们也可以为用户提供个性化、定制化的学者数据。欢迎来电咨询!咨询电话:010-8811{复制后删除}0370

www.globalauthorid.com

TOP

Copyright © 2019-2024 北京同舟云网络信息技术有限公司
京公网安备11010802033243号  京ICP备18003416号-3