# Scores of generated queries 

This repo contains the scores files pertaining to [this study](https://github.com/Watheq9/d2qminus-repro). In particular, we scored the expansion queries generated by [T5-based Doc2Query model](https://huggingface.co/castorini/monot5-base-msmarco) for MSMARCO-v1 passage dataset and a subset of BEIR benchemark.
We used [ELECTRA](https://huggingface.co/crystina-z/monoELECTRA_LCE_nneg31) cross-encoder to get the relevance scores between the document text and its expansion queries. More details in the study repo [here](https://github.com/Watheq9/d2qminus-repro).

## Structure 
All files are .jsonl files with the following three columns per line: ["id", "predicted_queries","querygen_score"].
So, each file contains the document id, the expansion queries and their corresponding ELECTRA relevance scores.
Here are the matching of each dataset:

`msmarco-v1-80-scored-queries.jsonl` is for MSMarco-v1 dataset.

`dbpedia-20-scored-queries.jsonl` is for DBPedia dataset.

`quora-20-scored-queries.jsonl` is for Quora dataset.

`robust04-20-scored-queries.jsonl` is for Robust04 dataset.

`trec-covid-20-scored-queries.jsonl` is for TREC-COVID dataset.

`webis-touche2020-20-scored-queries.jsonl` is for Touché-2020 dataset.

## Credit 
The N=80 expansion queries of MSMARCO-v1 were copied from this [repository](https://github.com/castorini/docTTTTTquery). Please cite their work.

The N=20 expansion queries of BEIR benchemark were copied from this [repository](https://huggingface.co/income). Please cite their work.

## Citation

If you used any piece of this repository, please consider citing our work:

```plaintext
@inproceedings{mansour2024revisit,
  title={Revisiting Document Expansion and Filtering for Effective First-Stage Retrieval},
  author={Mansour, Watheq and Zhuang, Shengyao and Zhuang, Guido and Mackenzie, Joel},
  booktitle = {Proceedings of the 47th International ACM SIGIR Conference on Research and Development in Information Retrieval},
  year={2024},
  publisher = {Association for Computing Machinery},
  series = {SIGIR '24}
}
```


---
license: cc-by-4.0
---