Automatic symbolic compositional verification by learning assumptions

Abstract

Compositional reasoning aims to improve scalability of verification tools by reducing the original verification task into subproblems. The simplification is typically based on assume-guarantee reasoning principles, and requires user guidance to identify appropriate assumptions for components. In this paper, we propose a fully automated approach to compositional reasoning that consists of automated decomposition using a hypergraph partitioning algorithm for balanced clustering of variables, and discovering assumptions using the L * algorithm for active learning of regular languages. We present a symbolic implementation of the learning algorithm, and incorporate it in the model checker NUSMV. In some cases, our experiments demonstrate significant savings in the computational requirements of symbolic model checking.

Automated Learning of Probabilistic Assumptions for Compositional Reasoning

“…The work has sparked a significant amount of interest in the area in recent years. There have been several attempts to improve performance, including symbolic implementations [31] and optimisations to the use of L* [9]. Others have also devised alternative learning-based methods, for example by reformulating the assumption generation problem as one of computing the smallest finite automaton separating two regular languages [23,11], or using the CDNF learning algorithm to generate implicit representations of assumptions [10].…”

Symbolic assume-guarantee reasoning through BDD learning

“…The authors apply the L * algorithm to generate explicit deterministic finite automata as contextual assumptions [2]. Several optimizations for the L * algorithm are available [32,9,35,4]. A tree-based contextual assumption generation algorithm is also developed in [26,27].…”

“…The symbolic compositional verification technique in [1,32] is only relevant in appearance. Based on the L * learning algorithm, the symbolic L * algorithm uses BDD's to encode input symbols on transitions of contextual assumptions.…”

“…In [20,1,8,35,24,32,25,13,34], the L * learning algorithm for regular languages [2] is adopted to solve the assumption generation problem. Instead of programmers, one employs the L * learning algorithm to contrive contextual assumptions under the supervision of a mechanical teacher.…”

Assume-Guarantee Reasoning With Local Specifications

“…Nam, Alur et al [22,23] proposed a symbolic approach to learning-based assume-guarantee reasoning. Farzan, Chen et al [8] extended the assume-guarantee rules to liveness properties, based on the fact that ω-regular languages preserve the essential closure properties of regular languages.…”