Reproduce “TRACE-CS: A Hybrid Logic-LLM System for Explainable Course Scheduling” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Stylianos Loukas Vasileiou, William Yeoh
TRACE-CS presents a hybrid AI system combining symbolic reasoning with Large Language Models to generate trustworthy and accessible natural language explanations for course scheduling decisions. This system achieves 100% logical correctness in explanations while delivering them in natural language, addressing limitations of purely LLM-based approaches in formal reasoning.
Key insights
- Grounding explanations in symbolic logic is crucial for achieving 100% logical correctness and trustworthiness in complex constraint-satisfaction problems.
- Pure LLMs frequently generate incorrect and verbose explanations when tasked with precise logical reasoning, exhibiting an overall correctness of only 54.1% in the evaluated context.
- A hybrid approach effectively combines the strengths of symbolic AI for rigorous reasoning with LLMs' natural language generation capabilities, overcoming individual limitations.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.