Audit finds 36-39% incorrect FOL labels in FOLIO and MALLS; corrections raise LLM accuracy 9-22 points and an LLM-guided review framework achieves 90% dataset quality after checking fewer than 24% of examples.
Yanwei Wang, Nadia Figueroa, Shen Li, Ankit Shah, and Julie Shah
5 Pith papers cite this work. Polarity classification is still indexing.
5
Pith papers citing it
years
2026 5representative citing papers
The work formalizes and implements an LLM-based NL-to-LTL translation pipeline for abstraction-based controller design, introduces a benchmark, and reports that translation accuracy falls with increasing LTL complexity measures.
AeroReq2LTL automates LTL generation from industrial aerospace requirements via LLMs with a data dictionary and templates, achieving 85% precision and 88% recall on real data.
VeriTrans achieves 94.46% SAT/UNSAT correctness on SatBench via LLM translation gated by round-trip similarity and deterministic neuro-symbolic execution.
This survey presents temporal logics as expressive specification languages and outlines formal synthesis techniques from discrete to continuous methods for robot planning and control, while noting real-world deployment challenges.
citing papers explorer
-
Fixing FOLIO and MALLS: Verified Annotations and an LLM-assisted Framework to Focus Human Relabeling
Audit finds 36-39% incorrect FOL labels in FOLIO and MALLS; corrections raise LLM accuracy 9-22 points and an LLM-guided review framework achieves 90% dataset quality after checking fewer than 24% of examples.
-
End-to-End Abstraction-Based Control with LLM-Enhanced NL-to-LTL Translation
The work formalizes and implements an LLM-based NL-to-LTL translation pipeline for abstraction-based controller design, introduces a benchmark, and reports that translation accuracy falls with increasing LTL complexity measures.
-
Automated LTL Specification Generation from Industrial Aerospace Requirements
AeroReq2LTL automates LTL generation from industrial aerospace requirements via LLMs with a data dictionary and templates, achieving 85% precision and 88% recall on real data.
-
VeriTrans: Fine-Tuned LLM-Assisted NL-to-PL Translation via a Deterministic Neuro-Symbolic Pipeline
VeriTrans achieves 94.46% SAT/UNSAT correctness on SatBench via LLM translation gated by round-trip similarity and deterministic neuro-symbolic execution.
-
Temporal logics and formal synthesis for robot planning and control
This survey presents temporal logics as expressive specification languages and outlines formal synthesis techniques from discrete to continuous methods for robot planning and control, while noting real-world deployment challenges.