Reinforcement Learning-Based Evidence Selection for Trustworthy RAG in Financial Analysis and Regulatory Compliance
Keywords:
retrieval-augmented generation; reinforcement learning; evidence selection; financial analysis; regulatory compliance; trustworthy artificial intelligence; governance; robustness; fairness; socio-technical systemsAbstract
Financial analysis and regulatory compliance depend on evidence that is accurate, current, traceable, and defensible under scrutiny. Retrieval-augmented generation has emerged as a practical architecture for grounding language models in external documents, but its trustworthiness in high-stakes financial settings is limited by weak evidence selection, ambiguous provenance, and unstable behavior under distributional shift. This paper examines reinforcement learning-based evidence selection as a system-level mechanism for improving the reliability of retrieval-augmented generation in financial analysis and regulatory compliance. Rather than treating retrieval as a static preprocessing step, the paper frames evidence selection as a sequential decision problem in which an agent learns to choose, sequence, and reject retrieved passages according to task objectives, risk constraints, and governance requirements. The discussion emphasizes architectural trade-offs involving latency, auditability, human oversight, compute sustainability, and fairness. It also considers how reinforcement learning policies can be aligned with regulatory expectations through constrained rewards, uncertainty-aware stopping criteria, and transparent escalation pathways. Case illustrations from earnings analysis, anti-money laundering review, and regulatory reporting show that trustworthy retrieval-augmented generation requires more than improved ranking accuracy. It requires socio-technical infrastructure for provenance, contestability, monitoring, and policy enforcement. The paper concludes that reinforcement learning-based evidence selection can strengthen financial artificial intelligence only when embedded within governance frameworks that treat evidence quality, uncertainty, and accountability as first-class design concerns.
References
1. Lewis, P., Perez, E., Piktus, A., Petroni, F., Karpukhin, V., Goyal, N., Küttler, H., Lewis, M., Yih, W., Rocktäschel, T., Riedel, S., & Kiela, D. (2020). Retrieval-augmented generation for knowledge-intensive NLP tasks. In Advances in Neural Information Processing Systems (Vol. 33, pp. 9459-9474). Curran Associates.
2. Karpukhin, V., Oğuz, B., Min, S., Lewis, P., Wu, L., Edunov, S., Chen, D., & Yih, W. (2020). Dense passage retrieval for open-domain question answering. In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (pp. 6769-6781). Association for Computational Linguistics.
3. Guu, K., Lee, K., Tung, Z., Pasupat, P., & Chang, M. (2020). Retrieval augmented language model pre-training. In Proceedings of the 37th International Conference on Machine Learning (Vol. 119, pp. 3929-3938). PMLR.
4. Izacard, G., & Grave, E. (2021). Leveraging passage retrieval with generative models for open domain question answering. In Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics (pp. 874-880). Association for Computational Linguistics.
5. Borgeaud, S., Mensch, A., Hoffmann, J., Cai, T., Rutherford, E., Millican, K., van den Driessche, G., Lespiau, J.-B., Damoc, B., Clark, A., de Las Casas, D., Guy, A., Menick, J., Ring, R., Hennigan, T., Huang, S., Maggiore, L., Jones, C., Cassirer, A., ... Sifre, L. (2022). Improving language models by retrieving from trillions of tokens. In Proceedings of the 39th International Conference on Machine Learning (Vol. 162, pp. 2206-2240). PMLR.
6. Asai, A., Wu, Z., Wang, Y., Sil, A., & Hajishirzi, H. (2024). Self-RAG: Learning to retrieve, generate, and critique through self-reflection. In The Twelfth International Conference on Learning Representations. OpenReview.net.
7. Jiang, Z., Xu, F. F., Gao, L., Sun, Z., Liu, Q., Dwivedi-Yu, J., Yang, Y., Callan, J., & Neubig, G. (2023). Active retrieval augmented generation. In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing (pp. 7969-7992). Association for Computational Linguistics.
8. Gao, Y., Xiong, Y., Gao, X., Jia, K., Pan, J., Bi, Y., Dai, Y., Sun, J., & Wang, H. (2023). Retrieval-augmented generation for large language models: A survey. arXiv. https://arxiv.org/abs/2312.10997
9. Sutton, R. S., & Barto, A. G. (2018). Reinforcement learning: An introduction (2nd ed.). MIT Press.
10. Mnih, V., Kavukcuoglu, K., Silver, D., Rusu, A. A., Veness, J., Bellemare, M. G., Graves, A., Riedmiller, M., Fidjeland, A. K., Ostrovski, G., Petersen, S., Beattie, C., Sadik, A., Antonoglou, I., King, H., Kumaran, D., Wierstra, D., Legg, S., & Hassabis, D. (2015). Human-level control through deep reinforcement learning. Nature, 518(7540), 529-533.
11. Schulman, J., Wolski, F., Dhariwal, P., Radford, A., & Klimov, O. (2017). Proximal policy optimization algorithms. arXiv. https://arxiv.org/abs/1707.06347
12. Christiano, P. F., Leike, J., Brown, T., Martic, M., Legg, S., & Amodei, D. (2017). Deep reinforcement learning from human preferences. In Advances in Neural Information Processing Systems (Vol. 30, pp. 4299-4307). Curran Associates.
13. Ouyang, L., Wu, J., Jiang, X., Almeida, D., Wainwright, C., Mishkin, P., Zhang, C., Agarwal, S., Slama, K., Ray, A., Schulman, J., Hilton, J., Kelton, F., Miller, L., Simens, M., Askell, A., Welinder, P., Christiano, P., Leike, J., & Lowe, R. (2022). Training language models to follow instructions with human feedback. In Advances in Neural Information Processing Systems (Vol. 35, pp. 27730-27744). Curran Associates.
14. Ziegler, D. M., Stiennon, N., Wu, J., Brown, T. B., Radford, A., Amodei, D., Christiano, P., & Irving, G. (2019). Fine-tuning language models from human preferences. arXiv. https://arxiv.org/abs/1909.08593
15. Amodei, D., Olah, C., Steinhardt, J., Christiano, P., Schulman, J., & Mané, D. (2016). Concrete problems in AI safety. arXiv. https://arxiv.org/abs/1606.06565
16. Zhang, R. (2025, December). Uncertainty-Aware RAG: Trustworthy Generation and Risk Control. In 2025 5th International Symposium on Artificial Intelligence and Big Data (AIBDF) (pp. 526-531). IEEE.
17. Arner, D. W., Barberis, J. N., & Buckley, R. P. (2017). FinTech and RegTech: Enabling innovation while preserving financial stability. Georgetown Journal of International Affairs, 18(3), 47-58.
18. Buchak, G., Matvos, G., Piskorski, T., & Seru, A. (2018). Fintech, regulatory arbitrage, and the rise of shadow banks. Journal of Financial Economics, 130(3), 453-483.
19. Mehrabi, N., Morstatter, F., Saxena, N., Lerman, K., & Galstyan, A. (2021). A survey on bias and fairness in machine learning. ACM Computing Surveys, 54(6), Article 115.
20. Barocas, S., Hardt, M., & Narayanan, A. (2019). Fairness and machine learning: Limitations and opportunities. fairmlbook.org.
Downloads
Published
Issue
Section
License
Copyright (c) 2026 Journal of Engineering Systems and Digital Innovation
This work is licensed under a Creative Commons Attribution 4.0 International License.
