The Solution for The AQA-KDD-2024 OAG-Challenge

Zhe Zhang, Weili Guo, Yizhe Zhang

Keywords: academic question answering

Abstract:

This report presents our solution for the OAG-Challenge’s Academic Question Answering (AQA) task, which aims to retrieve relevant papers that answer specialized questions. We utilized vector
models to embed papers, simplified questions using large language
models (LLMs), and employed FAISS for vector-based retrieval. Finally, we applied a re-ranking model for detailed sorting of the
results. Our approach leverages pretrained LLMs for efficient and
effective text embedding and retrieval. The implementation details
and code are publicly available at https://github.com/qianlanzz/AQAKDD-2024.git

Submission Number: 32