@inproceedings{afa4589b98174109a378b511a05e04da,
title = "End-to-End AI Generated Runtime Verification from Natural Language Specification",
abstract = "We demonstrate how LLMs can be harnessed to synthesize runtime verification monitors from natural language specifications. We instruct the LLM through prompts to assemble the monitors through a sequence of stages. To start with, we employ the learned insights LLMs possess regarding linear temporal logic for translating natural language specifications into propositional past time LTL formulas. Next, we utilize a sequence of prompts to synthesize from the LTL specification a runtime verification monitor for the given formula. Part of these prompts enable the LLM to validate its own outputs, thereby significantly improving the probability of obtaining a correct monitor.",
author = "Itay Cohen and Doron Peled",
note = "Publisher Copyright: {\textcopyright} The Author(s) 2025.; 1st International Symposium on Leveraging Applications of Formal Methods, AISoLA 2023 ; Conference date: 23-10-2023 Through 28-10-2023",
year = "2025",
doi = "https://doi.org/10.1007/978-3-031-73741-1_23",
language = "الإنجليزيّة",
isbn = "9783031737404",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Science and Business Media Deutschland GmbH",
pages = "362--384",
editor = "Bernhard Steffen",
booktitle = "Bridging the Gap Between AI and Reality - 1st International Conference, AISoLA 2023, Selected Papers",
address = "ألمانيا",
}