Learning Formal Specifications from Membership and Preference Queries

TL;DR

This paper introduces a flexible active learning framework for formal specifications that combines membership queries and pair-wise preferences, enhancing robustness and applicability across domains.

Contribution

It proposes a novel framework integrating preference queries with membership labels for active specification learning, extending prior methods.

Findings

Improved robustness in specification learning

Effective across multiple domains

Enhanced flexibility over traditional methods

Abstract

Active learning is a well-studied approach to learning formal specifications, such as automata. In this work, we extend active specification learning by proposing a novel framework that strategically requests a combination of membership labels and pair-wise preferences, a popular alternative to membership labels. The combination of pair-wise preferences and membership labels allows for a more flexible approach to active specification learning, which previously relied on membership labels only. We instantiate our framework in two different domains, demonstrating the generality of our approach. Our results suggest that learning from both modalities allows us to robustly and conveniently identify specifications via membership and preferences.