Reliable Generation of Formal Specifications using Large Language Models
| dc.contributor.author | Kogler, Philipp | |
| dc.contributor.author | Falkner, Andreas | |
| dc.contributor.author | Sperl, Simon | |
| dc.contributor.editor | Dhungana, Deepak | |
| dc.contributor.editor | Lambers, Leen | |
| dc.contributor.editor | Bonorden, Leif | |
| dc.contributor.editor | Henning, Sören | |
| dc.date.accessioned | 2024-02-14T05:22:29Z | |
| dc.date.available | 2024-02-14T05:22:29Z | |
| dc.date.issued | 2024 | |
| dc.description.abstract | Recent pre-trained Large Language Models (LLMs) have demonstrated promising Natural Language Processing (NLP) and code generation abilities. However, the intrinsically unreliable output due to the probabilistic nature of LLMs imposes a major challenge as validity can generally not be guaranteed, making subsequent processing prone to errors. When LLMs are used to translate natural-language specifications to formal specifications, this limitation becomes evident. We propose a framework involving prompting and algorithmic post-processing that continuously interacts with the LLM to ensure strict syntactic validity and reasonable content correctness. Furthermore, we introduce a use-case in the domain of engineering processes for railway infrastructure and demonstrate that our approach is sufficiently mature for implementation in an industrial environment. | en |
| dc.identifier.doi | 10.18420/sw2024-ws_10 | |
| dc.identifier.uri | https://dl.gi.de/handle/20.500.12116/43507 | |
| dc.language.iso | en | |
| dc.pubPlace | Bonn | |
| dc.publisher | Gesellschaft für Informatik e.V. | |
| dc.relation.ispartof | SE 2024 - Companion | |
| dc.subject | Generative AI | |
| dc.subject | Large Language Models | |
| dc.subject | Reliable Code Generation | |
| dc.subject | Post-processing | |
| dc.subject | Domain-specific Languages | |
| dc.subject | Engineering Processes | |
| dc.title | Reliable Generation of Formal Specifications using Large Language Models | en |
| dc.type | Text/Conference Paper | |
| gi.citation.endPage | 153 | |
| gi.citation.startPage | 141 | |
| gi.conference.date | 26.- 27. Februar | |
| gi.conference.location | Linz | |
| gi.conference.sessiontitle | GENSE |
Dateien
Originalbündel
1 - 1 von 1
Vorschaubild nicht verfügbar
- Name:
- A4-1.pdf
- Größe:
- 246.1 KB
- Format:
- Adobe Portable Document Format