What is Formal Verification without Specifications? A Survey on mining LTL Specifications
Daniel Neider, Rajarshi Roy
TL;DR
The paper addresses the bottleneck of obtaining formal specifications for verification by surveying methods that mine Linear Temporal Logic (LTL) specifications from system executions. It categorizes approaches into constraint-based, enumeration-based, and neural-network-based, analyzes their guarantees, scalability, and applicability, and discusses extensions to noisy data, positive-only settings, templates, natural language inputs, and logics beyond LTL. The survey builds from foundational preliminaries on words and LTL syntax/semantics to a comprehensive review of practical algorithms, including SAT/MaxSAT encodings, ASP/ILP variants, GPU-accelerated enumeration, and graph neural network techniques, illustrating trade-offs between minimality, consistency, and interpretability. It offers a roadmap for practitioners and researchers, highlighting open challenges in integrating domain knowledge through templates, improving NL-to-LTL translation, and extending specification learning to STL, CTL, ATL, and related logics for broader applicability in reactive and cyber-physical systems.
Abstract
Virtually all verification techniques using formal methods rely on the availability of a formal specification, which describes the design requirements precisely. However, formulating specifications remains a manual task that is notoriously challenging and error-prone. To address this bottleneck in formal verification, recent research has thus focussed on automatically generating specifications for formal verification from examples of (desired and undesired) system behavior. In this survey, we list and compare recent advances in mining specifications in Linear Temporal Logic (LTL), the de facto standard specification language for reactive systems. Several approaches have been designed for learning LTL formulas, which address different aspects and settings of specification design. Moreover, the approaches rely on a diverse range of techniques such as constraint solving, neural network training, enumerative search, etc. We survey the current state-of-the-art techniques and compare them for the convenience of the formal methods practitioners.