This repository contains the code for the TRACE-CS framework, presented in the AAAI 2025 demo paper: Trustworthy Reasoning for Contrastive Explanations in Course Scheduling Problems.

TRACE-CS is a hybrid system that combines symbolic reasoning and large language models (LLMs) to provide contrastive explanations in course scheduling. By leveraging SAT-solving techniques and the natural language generation capabilities of LLMs, TRACE-CS ensures:

• Accurate and provably correct explanations.
• Linguistically coherent and user-friendly communication.
• Support for real-world course scheduling scenarios.

• Symbolic Module: Encodes scheduling constraints and generates logic-based explanations.
• LLM Module: Converts user queries to symbolic representations and refines explanations into natural language.

Clone the repository:

git clone https://github.com/YODA-Lab/TRACE-CS/

Navigate to the project directory:

cd trace-cs

Install dependencies:

pip install -r requirements.txt

To launch the GUI, simply run:

python main.py

If you use TRACE-CS, or you are inspired by TRACE-CS, please cite the paper:

@inproceedings{trace-cs2025,
  title={Trustworthy Reasoning for Contrastive Explanations in Course Scheduling Problems},
  author={Stylianos Loukas Vasileiou, William Yeoh},
  booktitle={Proceedings of the AAAI Conference on Artificial Intelligence},
  year={2025}
}

This project is licensed under the MIT License.