Abstract
Retrieval Augmented Generation (RAG) has become a common practice to alleviate the hallucination of Large Language Models (LLMs). The retrieval phase of RAG, however, usually solely depends on the original query, which, to some extent, suffers from the problem of semantic gap and thus degrades the quality of the retrieved external knowledge. To address this problem and enhance the performance of the traditional RAG, we propose a rEwrite-sElect-votE-rEad paradigm (
) that first paraphrases the original query into N rewritten ones to bridge the semantic gap from different perspectives and then determines the most valuable retrieved external knowledge via a voting manner. Besides, in the midst of the above procedures, a certain query-selecting strategy is also required to filter out the extra noise introduced by the query-rewriting process. Following this proposed paradigm, we provide our implementation of
. Experimental results of our implementation on long context reading comprehension datasets from LongBench demonstrate the effectiveness of our proposed paradigm and provide a profound insight into the whole enhanced RAG process.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Similar content being viewed by others
Notes
- 1.
Methods from information retrieval domain such as sparse and dense retrieval methods are widely used in this step.
- 2.
The long passage of a data sample in LCRC task can be regarded as the external non-parameterized knowledge and the question as the original query.
- 3.
Here, the best results among eight experimental settings are reported. In each column, the best result is in bold and results better than baselines are marked with
.
- 4.
In each row, the best result is in bold and results better than baseline are marked with
.
- 5.
Mean values of results under eight experimental settings are reported in Table 4.
References
Chen, T., et al.: Dense x retrieval: what retrieval granularity should we use? ArXiv abs/2312.06648 (2023)
Devlin, J., Chang, M.W., Lee, K., Toutanova, K.: BERT: pre-training of deep bidirectional transformers for language understanding. In: Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long and Short Papers), pp. 4171–4186 (2019)
Formal, T., Lassance, C., Piwowarski, B., Clinchant, S.: SPLADE v2: sparse lexical and expansion model for information retrieval. ArXiv abs/2109.10086 (2021)
Gao, L., Ma, X., Lin, J., Callan, J.: Precise zero-shot dense retrieval without relevance labels. In: Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pp. 1762–1777 (2023)
Gao, T., Yao, X., Chen, D.: SimCSE: simple contrastive learning of sentence embeddings. In: Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing (EMNLP), pp. 6894–6910 (2021)
Gao, Y., et al.: Retrieval-augmented generation for large language models: a survey. ArXiv abs/2312.10997 (2023)
Hofstätter, S., Lin, S.C., Yang, J.H., Lin, J., Hanbury, A.: Efficiently teaching an effective dense retriever with balanced topic aware sampling. In: Proceedings of the 44th International ACM SIGIR Conference on Research and Development in Information Retrieval, pp. 113–122 (2021)
Huang, L., et al.: A survey on hallucination in large language models: principles, taxonomy, challenges, and open questions. ArXiv abs/2311.05232 (2023)
Izacard, G., et al.: Unsupervised dense information retrieval with contrastive learning. ArXiv abs/2112.09118 (2021)
Karpukhin, V., et al.: Dense passage retrieval for open-domain question answering. In: Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP), pp. 6769–6781 (2020)
Liu, N.F., et al.: Lost in the middle: how language models use long contexts. Trans. Assoc. Comput. Linguist. 12, 157–173 (2024)
Liu, Y., et al.: RoBERTa: a robustly optimized BERT pretraining approach. ArXiv abs/1907.11692 (2019)
Lu, Y., Bartolo, M., Moore, A., Riedel, S., Stenetorp, P.: Fantastically ordered prompts and where to find them: overcoming few-shot prompt order sensitivity. In: Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pp. 8086–8098 (2022)
Ma, X., Gong, Y., He, P., Zhao, H., Duan, N.: Query rewriting in retrieval-augmented large language models. In: Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing (EMNLP), pp. 5303–5315 (2023)
Mitra, M., Chaudhuri, B.: Information retrieval from documents: a survey. Inf. Retrieval 2, 141–163 (2000)
Muresanu, A., Thudi, A., Zhang, M.R., Papernot, N.: Unlearnable algorithms for in-context learning. ArXiv abs/2402.00751 (2024)
Nussbaum, Z., Morris, J.X., Duderstadt, B., Mulyar, A.: Nomic embed: training a reproducible long context text embedder. ArXiv abs/2402.01613 (2024)
Peng, W., et al.: Large language model based long-tail query rewriting in taobao search. In: Companion Proceedings of the ACM on Web Conference 2024, pp. 20–28 (2024)
Robertson, S., Zaragoza, H.: The probabilistic relevance framework: BM25 and beyond. Inf. Retrieval 3(4), 333–389 (2009)
Sun, Y., et al.: ERNIE: enhanced representation through knowledge integration. ArXiv abs/1904.09223 (2019)
Touvron, H., et al.: LLaMA 2: open foundation and fine-tuned chat models. ArXiv abs/2307.09288 (2023)
Vaswani, A., et al.: Attention is all you need. Advances in Neural Information Processing Systems, vol. 30 (2017)
Wang, Y., et al.: Self-instruct: Aligning language models with self-generated instructions. In: Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pp. 13484–13508 (2023)
Wei, J., et al.: Emergent abilities of large language models. ArXiv abs/2206.07682 (2022)
Wu, T., et al.: A brief overview of ChatGPT: the history, status quo and potential future development. IEEE/CAA J. Autom. Sinica 10(5), 1122–1136 (2023)
Zhao, W.X., Liu, J., Ren, R., Wen, J.R.: Dense text retrieval based on pretrained language models: A survey. ACM Trans. Inf. Syst. 42(4), 1–60 (2024)
Zhu, Y., et al.: Large language models for information retrieval: a survey. ArXiv abs/2308.07107 (2023)
Acknowledgments
This work is supported by the Youth Innovation Promotion Association of the Chinese Academy of Sciences (E1291902), Jun Zhou (2021025).
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Ethics declarations
Disclosure of Interests
The authors have no competing interests to declare that are relevant to the content of this paper.
Rights and permissions
Copyright information
© 2025 The Author(s), under exclusive license to Springer Nature Switzerland AG
About this paper
Cite this paper
Guan, W., Li, X., Lu, J., Zhou, J. (2025).
: A Voting-Based Paradigm for Enhancing Retrieval Augmented Generation.
In: Antonacopoulos, A., Chaudhuri, S., Chellappa, R., Liu, CL., Bhattacharya, S., Pal, U. (eds) Pattern Recognition. ICPR 2024. Lecture Notes in Computer Science, vol 15331. Springer, Cham. https://doi.org/10.1007/978-3-031-78119-3_9
Download citation
DOI: https://doi.org/10.1007/978-3-031-78119-3_9
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-031-78118-6
Online ISBN: 978-3-031-78119-3
eBook Packages: Computer ScienceComputer Science (R0)