# FinCPRG: A Bidirectional Generation Pipeline for Hierarchical Queries and Rich Relevance in Financial Chinese Passage Retrieval

arXiv: https://arxiv.org/abs/2508.02222 · HTML: https://mlsanity.com/paper/2508.02222/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 21 of its 28 references that are in the map.
- First posted on arXiv 4 Aug 2025.

## Builds on (most cited first, 21 of 21)

- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [BEIR: A Heterogenous Benchmark for Zero-shot Evaluation of Information Retrieval Models](https://mlsanity.com/paper/2104.08663/) (2021)
- [It's Not Just Size That Matters: Small Language Models Are Also Few-Shot Learners](https://mlsanity.com/paper/2009.07118/) (2020)
- [MTEB: Massive Text Embedding Benchmark](https://mlsanity.com/paper/2210.07316/) (2022)
- [Precise Zero-Shot Dense Retrieval without Relevance Labels](https://mlsanity.com/paper/2212.10496/) (2022)
- [Improving Text Embeddings with Large Language Models](https://mlsanity.com/paper/2401.00368/) (2023)
- [Time2Vec: Learning a Vector Representation of Time](https://mlsanity.com/paper/1907.05321/) (2019)
- [Query2doc: Query Expansion with Large Language Models](https://mlsanity.com/paper/2303.07678/) (2023)
- [GPL: Generative Pseudo Labeling for Unsupervised Domain Adaptation of Dense Retrieval](https://mlsanity.com/paper/2112.07577/) (2021)
- [InPars-v2: Large Language Models as Efficient Dataset Generators for Information Retrieval](https://mlsanity.com/paper/2301.01820/) (2023)
- [T2Ranking: A large-scale Chinese Benchmark for Passage Ranking](https://mlsanity.com/paper/2304.03679/) (2023)
- [FinTextQA: A Dataset for Long-form Financial Question Answering](https://mlsanity.com/paper/2405.09980/) (2024)
- [When do Generative Query and Document Expansions Fail? A Comprehensive Study Across Methods, Retrievers, and Datasets](https://mlsanity.com/paper/2309.08541/) (2023)
- [Contrastive Learning of Sentence Embeddings from Scratch](https://mlsanity.com/paper/2305.15077/) (2023)
- [Synthetic Test Collections for Retrieval Evaluation](https://mlsanity.com/paper/2405.07767/) (2024)
- [FinMTEB: Finance Massive Text Embedding Benchmark](https://mlsanity.com/paper/2502.10990/) (2025)
- [Cocktail: A Comprehensive Information Retrieval Benchmark with LLM-Generated Documents Integration](https://mlsanity.com/paper/2405.16546/) (2024)
- [DuReader_retrieval: A Large-scale Chinese Benchmark for Passage Retrieval from Web Search Engine](https://mlsanity.com/paper/2203.10232/) (2022)
- [AIR-Bench: Automated Heterogeneous Information Retrieval Benchmark](https://mlsanity.com/paper/2412.13102/) (2024)
- [CSPRD: A Financial Policy Retrieval Dataset for Chinese Stock Market](https://mlsanity.com/paper/2309.04389/) (2023)
- [FinBERT2: A Specialized Bidirectional Encoder for Bridging the Gap in Finance-Specific Deployment of Large Language Models](https://mlsanity.com/paper/2506.06335/) (2025)

## Abstract (arXiv)

In recent years, large language models (LLMs) have demonstrated significant potential in constructing passage retrieval datasets. However, existing methods still face limitations in expressing cross-doc query needs and controlling annotation quality. To address these issues, this paper proposes a bidirectional generation pipeline, which aims to generate 3-level hierarchical queries for both intra-doc and cross-doc scenarios and mine additional relevance labels on top of direct mapping annotation. The pipeline introduces two query generation methods: bottom-up from single-doc text and top-down from multi-doc titles. The bottom-up method uses LLMs to disassemble and generate structured queries at both sentence-level and passage-level simultaneously from intra-doc passages. The top-down approach incorporates three key financial elements--industry, topic, and time--to divide report titles into clusters and prompts LLMs to generate topic-level queries from each cluster. For relevance annotation, our pipeline not only relies on direct mapping annotation from the generation relationship but also implements an indirect positives mining method to enrich the relevant query-passage pairs. Using this pipeline, we constructed a Financial Passage Retrieval Generated dataset (FinCPRG) from almost 1.3k Chinese financial research reports, which includes hierarchical queries and rich relevance labels. Through evaluations of mined relevance labels, benchmarking and training experiments, we assessed the quality of FinCPRG and validated its effectiveness as a passage retrieval dataset for both training and benchmarking.