PCPs and Instance Compression from a Cryptographic Lens

Liron Bronfman, Ron D. Rothblum

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Modern cryptography fundamentally relies on the assumption that the adversary trying to break the scheme is computationally bounded. This assumption lets us construct cryptographic protocols and primitives that are known to be impossible otherwise. In this work we explore the effect of bounding the adversary's power in other information theoretic proof-systems and show how to use this assumption to bypass impossibility results. We first consider the question of constructing succinct PCPs. These are PCPs whose length is polynomial only in the length of the original NP witness (in contrast to standard PCPs whose length is proportional to the non-deterministic verification time). Unfortunately, succinct PCPs are known to be impossible to construct under standard complexity assumptions. Assuming the sub-exponential hardness of the learning with errors (LWE) problem, we construct succinct probabilistically checkable arguments or PCAs (Kalai and Raz 2009), which are PCPs in which soundness is guaranteed against efficiently generated false proofs. Our PCA construction is for every NP relation that can be verified by a small-depth circuit (e.g., SAT, clique, TSP, etc.) and in contrast to prior work is publicly verifiable and has constant query complexity. Curiously, we also show, as a proof-of-concept, that such publicly-verifiable PCAs can be used to derive hardness of approximation results. Second, we consider the notion of Instance Compression (Harnik and Naor, 2006). An instance compression scheme lets one compress, for example, a CNF formula φ on m variables and n ≫ m clauses to a new formula φ with only poly(m) clauses, so that φ is satisfiable if and only if φ is satisfiable. Instance compression has been shown to be closely related to succinct PCPs and is similarly highly unlikely to exist. We introduce a computational analog of instance compression in which we require that if φ is unsatisfiable then φ is effectively unsatisfiable, in the sense that it is computationally infeasible to find a satisfying assignment for φ (although such an assignment may exist). Assuming the same sub-exponential LWE assumption, we construct such computational instance compression schemes for every bounded-depth NP relation. As an application, this lets one compress k formulas ϕ1,..., ϕk into a single short formula ϕ that is effectively satisfiable if and only if at least one of the original formulas was satisfiable.

Original languageEnglish
Title of host publication13th Innovations in Theoretical Computer Science Conference, ITCS 2022
EditorsMark Braverman
ISBN (Electronic)9783959772174
DOIs
StatePublished - 1 Jan 2022
Event13th Innovations in Theoretical Computer Science Conference, ITCS 2022 - Berkeley, United States
Duration: 31 Jan 20223 Feb 2022

Publication series

NameLeibniz International Proceedings in Informatics, LIPIcs
Volume215

Conference

Conference13th Innovations in Theoretical Computer Science Conference, ITCS 2022
Country/TerritoryUnited States
CityBerkeley
Period31/01/223/02/22

Keywords

  • Instance compression
  • PCP
  • Succinct arguments

All Science Journal Classification (ASJC) codes

  • Software

Fingerprint

Dive into the research topics of 'PCPs and Instance Compression from a Cryptographic Lens'. Together they form a unique fingerprint.

Cite this