English

Auditable Algorithms for Approximate Model Counting

Logic in Computer Science 2023-12-20 v1 Artificial Intelligence

Abstract

Model counting, or counting the satisfying assignments of a Boolean formula, is a fundamental problem with diverse applications. Given #P-hardness of the problem, developing algorithms for approximate counting is an important research area. Building on the practical success of SAT-solvers, the focus has recently shifted from theory to practical implementations of approximate counting algorithms. This has brought to focus new challenges, such as the design of auditable approximate counters that not only provide an approximation of the model count, but also a certificate that a verifier with limited computational power can use to check if the count is indeed within the promised bounds of approximation. Towards generating certificates, we start by examining the best-known deterministic approximate counting algorithm that uses polynomially many calls to a Σ2P\Sigma_2^P oracle. We show that this can be audited via a Σ2P\Sigma_2^P oracle with the query constructed over n2log2nn^2 \log^2 n variables, where the original formula has nn variables. Since nn is often large, we ask if the count of variables in the certificate can be reduced -- a crucial question for potential implementation. We show that this is indeed possible with a tradeoff in the counting algorithm's complexity. Specifically, we develop new deterministic approximate counting algorithms that invoke a Σ3P\Sigma_3^P oracle, but can be certified using a Σ2P\Sigma_2^P oracle using certificates on far fewer variables: our final algorithm uses only nlognn \log n variables. Our study demonstrates that one can simplify auditing significantly if we allow the counting algorithm to access a slightly more powerful oracle. This shows for the first time how audit complexity can be traded for complexity of approximate counting.

Keywords

Cite

@article{arxiv.2312.12362,
  title  = {Auditable Algorithms for Approximate Model Counting},
  author = {Kuldeep S. Meel and Supratik Chakraborty and S. Akshay},
  journal= {arXiv preprint arXiv:2312.12362},
  year   = {2023}
}

Comments

Full version of conference paper accepted at AAAI'24. The authors decided to forgo the old convention of alphabetical ordering of authors in favor of a randomized ordering. The publicly verifiable record of the randomization is available at https://www.aeaweb.org/journals/policies/random-author-order/search