Phase Description Deadline Download
Phase 1 Build a search engine for document retrieval: implement an indexer, preprocess queries, parse boolean queries, and retrieve relevant documents. Use the provided autograder to validate correctness. Ordibehesht 23 , 1405 📄 ZIP
Phase 2 Complete IR/ML pipeline on an IMDB movie dataset. Build inverted index, TF‑IDF matrix, feature selection, train/evaluate classifiers (Rocchio, k‑NN, Linear SVM) and clustering, analyze effect of feature reduction. Khordad 6 , 1405 📄 ZIP
Phase 3 Neural Retrieval and Retrieval‑Augmented Generation (RAG) on IMDB data. Implement embedding generation, FAISS vector index, query expansion, RAG chain, and evaluate retrieval quality using a Jupyter notebook. Khordad 27 , 1405 📄 ZIP

📌 Note: Detailed instructions and submission guidelines will be provided via the Telegram channel. The autograder and Jupyter notebooks contain specific TODO sections. Make sure to restart and run all cells before final submission. For any download issues, contact the teaching assistants.