Query Expansion & Re-Ranking for Biomedical Information Retrieval

Overview

This project explores advanced query expansion and document re-ranking techniques to improve biomedical information retrieval. It was developed as part of a bachelor thesis and focuses on enhancing search accuracy in large-scale biomedical datasets, particularly the TREC-COVID dataset.

Key Features

• Query Expansion: Implements synonym-based query expansion using Word2Vec and Semantic Knowledge Graphs (SKG).
• Re-Ranking with Transformers: Utilizes BERT-based sentence embeddings to refine search results.
• Evaluation & Benchmarking: Compares methods against the BM25 baseline, showing improved precision and recall.
• Scalable Search Implementation: Uses Elasticsearch for document indexing and retrieval.

Thesis Contribution

• Proposes novel query expansion methods that dynamically enhance search queries.
• Integrates neural ranking models for effective document re-ranking.
• Demonstrates state-of-the-art results on the TREC-COVID dataset, outperforming standard retrieval methods.

Folder Structure

📂 Thesis
 ├── 📜 code/                   # Code implementation
 ├── 📜 results/               # Experimental results & benchmarks
 ├── 📜 README.md              # Project documentation (this file)
 └── 📜 thesis.pdf             # Full thesis document

**The Whole TREC_COVID Dataset Can Be Found Through The Official Site Of "National Institute of Standards and Technology" **

https://ir.nist.gov/trec-covid/data.html

Citation

If you use this project in your work, please cite:
Ioannidis, V. (2024). "Query Expansion Techniques and Document Re-Ranking for Biomedical Information Retrieval."

---