A Birthday Repetition Theorem and Complexity of Approximating Dense CSPs

Abstract

A (k× l)-birthday repetition G of a two-prover game G is a game in which the two provers are sent random sets of questions from G of sizes k and l respectively. These two sets are sampled independently uniformly among all sets of questions of those particular sizes. We prove the following birthday repetition theorem: when G satisfies some mild conditions, val(G) decreases exponentially in Ω(kl/n) where n is the total number of questions. Our result positively resolves an open question posted by Aaronson, Impagliazzo and Moshkovitz [AIM14]. As an application of our birthday repetition theorem, we obtain new fine-grained hardness of approximation results for dense CSPs. Specifically, we establish a tight trade-off between running time and approximation ratio for dense CSPs by showing conditional lower bounds, integrality gaps and approximation algorithms. In particular, for any sufficiently large i and for every k ≥ 2, we show the following results: • We exhibit an O(q)-approximation algorithm for dense Max k-CSPs with alphabet size q via Ok(i)-level of Sherali-Adams relaxation. • Through our birthday repetition theorem, we obtain an integrality gap of q for Ω̃k(i)-level Lasserre relaxation for fully-dense Max k-CSP. • Assuming that there is a constant ε > 0 such that Max 3SAT cannot be approximated to within (1 − ε) of the optimal in sub-exponential time, our birthday repetition theorem implies that any algorithm that approximates fully-dense Max k-CSP to within a q factor takes (nq)k time, almost tightly matching the algorithmic result based on Sherali-Adams relaxation. As a corollary of our approximation algorithm for dense Max k-CSP, we give a new approximation algorithm for Densest k-Subhypergraph, a generalization of Densest k-Subgraph to hypergraphs. In particular, when the input hypergraph is O(1)-uniform and the optimal k-subhypergraph has constant density, our algorithm finds a k-subhypergraph of density Ω(n) in time n for any integer i > 0. University of California, Berkeley. Email: pasin@berkeley.edu. University of California, Berkeley. Email: prasad@berkeley.edu.

DOI: 10.4230/LIPIcs.ICALP.2017.78

Extracted Key Phrases

1 Figure or Table

0102020162017
Citations per Year

Citation Velocity: 7

Averaging 7 citations per year over the last 2 years.

Learn more about how we calculate this metric in our FAQ.

Cite this paper

@inproceedings{Manurangsi2017ABR, title={A Birthday Repetition Theorem and Complexity of Approximating Dense CSPs}, author={Pasin Manurangsi and Prasad Raghavendra}, booktitle={ICALP}, year={2017} }