Conflict Extraction in Probabilistic Datalog Analyses

2026-08-11

AI Summary

This work addresses a critical limitation in probabilistic Datalog-based program analysis: while individual alarms are often sound, they may be mutually exclusive, leading developers to investigate combinations that cannot co-occur in practice. To tackle this issue, the paper introduces PPProbe, the first conflict extractor specifically designed for this setting. PPProbe formalizes inconsistencies as minimal unsatisfiable subsets (MUSes) and innovatively integrates structural information from Datalog derivation graphs with probabilistic weights. It employs a structure-aware, bottom-up search strategy enhanced by UNSAT-core-based pruning to efficiently enumerate conflicts. Experimental evaluation on 70 benchmarks demonstrates that PPProbe achieves 2.5× to 24× higher throughput than the state-of-the-art MUS enumerator and, on average, filters out 47.7% of mutually exclusive alarms.

Abstract

Probabilistic extensions of Datalog enable static analyses such as pointer analysis, data race detection, and side-channel analysis to rank alarms by likelihood, but this added expressiveness also introduces a new challenge absent from deterministic analyses: the final output may contain alarms that are individually plausible yet mutually inconsistent, because marginal probabilities do not guarantee joint satisfiability. As a result, developers may spend effort investigating combinations of alarms that can never co-occur in any possible world. We address this problem by formalizing such inconsistencies as minimal unsatisfiable subsets (MUSes) and introducing PPProbe, a conflict extractor specialized for probabilistic Datalog analyses. Rather than improving MUS enumeration in general, PPProbe exploits the structure of Datalog derivation graphs to guide the search toward likely conflicts and prune the search space through bottom-up UNSAT inference. We evaluate PPProbe on 70 benchmarks from power side-channel analysis, data race detection, semantic diffing, and Bayesian-network inference. The results show that PPProbe achieves 2.5 to 24 times higher throughput than state-of-the-art MUS enumerators, and that the conflicts it identifies yield a conservative estimate of false-positive reduction, filtering out an average of 47.7% of mutually inconsistent alarms.

Problem

• Probabilistic Datalog
• conflict extraction
• mutual inconsistency
• minimal unsatisfiable subsets
• alarm ranking

Innovation

• Probabilistic Datalog
• Minimal Unsatisfiable Subsets
• Conflict Extraction
• Derivation Graph
• UNSAT Inference