📄 ddro-docids

This repository provides the generated document IDs (DocIDs) used for training and evaluating the DDRO (Direct Document Relevance Optimization) models.

Two types of DocIDs are included:

• PQ (Product Quantization) DocIDs: Compact semantic representations based on quantized document embeddings.
• TU (Title + URL) DocIDs: Tokenized document identifiers constructed from document titles and/or URLs.

---

📚 Contents

• pq_msmarco_docids.txt: PQ DocIDs for MS MARCO (MS300K).
• tu_msmarco_docids.txt: TU DocIDs for MS MARCO (MS300K).
• pq_nq_docids.txt: PQ DocIDs for Natural Questions (NQ320K).
• tu_nq_docids.txt: TU DocIDs for Natural Questions (NQ320K).

Each file maps a document ID to its corresponding tokenized docid representation.

---

📌 Details

• Maximum Length:
  • PQ DocIDs: Up to 24 tokens (24 subspaces for PQ coding).
  • TU DocIDs: Up to 99 tokens (after tokenization and truncation).
• Tokenizer and Model:
  • T5-Base tokenizer and model are used for tokenization.
• DocID Construction:
  • PQ DocIDs: Generated by quantizing dense document embeddings obtained from a SentenceTransformer (GTR-T5-Base) model.
  • TU DocIDs: Generated by tokenizing reversed URL segments or document titles combined with domains based on semantic richness.
• Final Adjustment:
  • All DocIDs are appended with [1] (end-of-sequence) token for consistent decoding.

---

🛠️ Code for Embedding and DocID Generation

Step 1: Generate Document Embeddings

Document embeddings are generated using a SentenceTransformer model (gtr-t5-base by default). The scripts used to generate these embbedings are available here.

Example:

python generate_embeddings.py \
    --input_path path/to/input.jsonl \
    --output_path path/to/save/embeddings.txt \
    --model_name sentence-transformers/gtr-t5-base \
    --batch_size 128 \
    --dataset msmarco

• input_path: Path to the document corpus.
• output_path: Destination for the generated embeddings.
• dataset: Choose msmarco or nq.

Note: For NQ, documents are loaded differently (from gzipped TSV format).

---

🛠️ Code for DocID Generation

The scripts used to generate these DocIDs are available here.

Key functionality:

• Loading document embeddings and documents.
• Encoding document IDs with PQ codes or URL/title-based tokenization.
• Applying consistent token indexing for training generative retrievers.

Example usage:

python generate_encoded_docids.py \
    --encoding pq \
    --input_doc_path path/to/documents.jsonl \
    --input_embed_path path/to/embeddings.txt \
    --output_path path/to/save/pq_docids.txt \
    --pretrain_model_path transformer_models/t5-base

Supported encoding options: atomic, pq, url, summary

📖 Citation

If you use these docids, please cite:

@article{mekonnen2025lightweight,
author = {Mekonnen, Kidist Amde and Tang, Yubao and de Rijke, Maarten},
title = {Lightweight and Direct Document Relevance Optimization for Generative Information Retrieval},
year = {2025},
isbn = {9798400715921},
publisher = {Association for Computing Machinery},
address = {New York, NY, USA},
url = {https://doi.org/10.1145/3726302.3730023},
doi = {10.1145/3726302.3730023},
abstract = {Generative information retrieval (GenIR) is a promising neural retrieval paradigm that formulates document retrieval as a document identifier (docid) generation task, allowing for end-to-end optimization toward a unified global retrieval objective. However, existing GenIR models suffer from token-level misalignment, where models trained to predict the next token often fail to capture document-level relevance effectively. While reinforcement learning-based methods, such as reinforcement learning from relevance feedback (RLRF), aim to address this misalignment through reward modeling, they introduce significant complexity, requiring the optimization of an auxiliary reward function followed by reinforcement fine-tuning, which is computationally expensive and often unstable. To address these challenges, we propose direct document relevance optimization (DDRO), which aligns token-level docid generation with document-level relevance estimation through direct optimization via pairwise ranking, eliminating the need for explicit reward modeling and reinforcement learning. Experimental results on benchmark datasets, including MS MARCO document and Natural Questions, show that DDRO outperforms reinforcement learning-based methods, achieving a 7.4\% improvement in MRR@10 for MS MARCO and a 19.9\% improvement for Natural Questions. These findings highlight DDRO's potential to enhance retrieval effectiveness with a simplified optimization approach. By framing alignment as a direct optimization problem, DDRO simplifies the ranking optimization pipeline of GenIR models while offering a viable alternative to reinforcement learning-based methods},
booktitle = {Proceedings of the 48th International ACM SIGIR Conference on Research and Development in Information Retrieval},
pages = {1327–1338},
numpages = {12},
keywords = {document relevance optimization, generative information retrieval, learning to rank, ranking optimization, supervised fine-tuning},
location = {Padua, Italy},
series = {SIGIR '25}
}

---