Adaptive and Enhanced Retrieval Augmented Generation (RAG) Systems: A Summarised Survey

Pit Pichappan

Pit Pichappan

Abstract

This brief review examines adaptive and enhanced Retrieval-Augmented Generation (RAG) systems, focusing on overcoming the limitations of standard RAG models, such as inefficiency, excessive resource utilisation, and rigid retrieve-then-generate workflows. It highlights advancements like Dynamic RAG and Parametric RAG, which enable context-aware retrieval and parameter-level knowledge integration. The paper emphasises adaptive mechanisms that intelligently decide when and how much to retrieve, improving efficiency and relevance. It also discusses query rewriting, verification, and multimodal extensions to enhance precision. Furthermore, the integration of forecasting latent retrieval methods is introduced, where deep latent dynamics models extract predictable, interpretable components from limited time series data, improving forecasting accuracy. Applications in education, edge computing, and domain-specific contexts are explored, showing reduced hallucinations and better scalability. The review outlines the evolution of RAG toward intelligent, responsive systems using reinforcement learning and hybrid architectures. We identified several potential future research directions, including agentic workflows, multi-modal adaptation, and domainspecific RAG models, which promise more reliable, scalable, and context-aware generative AI applications across various fields.

References

[1] Lu, P., Dong, X., Zhou, Y., Cheng, L., Yuan, C., Mo, L. (2025). DOGR: Leveraging Document-Oriented Contra/stive Learning in Generative Retrieval, In: Proceedings of the AAAI Conference on Artificial Intelligence, 39 (23), 24732-24740 [2] Lee, Dohyeon., Kim, Jongyoon., Kim, Jihyuk., Hwang., Seung-won., Park, Joonsuk. (2025). tRAG: Termlevel retrieval-augmented generation for domain-adaptive retrieval. In Proceedings of the 2025 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers) (pp. 6566–6578). [3] Zheng, Xu., Weng, Ziqiao., Lyu, Yuanhuiyi., Jiang, Lutao., Xue, Haiwei., Ren, Bin., Paudel, Danda., Sebe, Nicu., Van Gool, Luc., Hu, Xuming. (2025). Retrieval-augmented generation and understanding in vision: A survey and new outlook. arXiv:2503.18016v1 [cs.CV] 23 Mar 2025 [4] Qin, Qitao., Luo, Yucong., Lu, Yihang., Chu, Zhibo., Liu, Xiaoman., Meng, Xianwei. (2025). Towards adaptive memory-based optimization for enhanced retrieval-augmented generation. arXiv:2504.05312v3 - March 2025. [5] Ouyang, T., et al. (2025). AdaRAG: Adaptive optimization for retrieval augmented generation with multilevel retrievers at the edge. In IEEE INFOCOM 2025 - IEEE Conference on Computer Communications (pp. 1–10). doi.org/10.1109/INFOCOM55648.2025.11044685 [6] Li, Siran., Stenzel, Linus., Eickhoff, Carsten., Bahrainian, Seyed Ali. (2025). Enhancing retrieval-augmented generation: A study of best practices. arXiv:2501.07391v1 [cs.CL] 13 Jan 2025]. [7] Shan, R. (2025). LearnRAG: Implementing retrieval-augmented generation for adaptive learning systems. In: 2025 International Conference on Artificial Intelligence in Information and Communication (ICAIIC) (pp. 224–229).[8] Su, Weihang., Ai, Qingyao., Zhan, Jingtao., Dong, Qian., Liu, Yiqun. (2025). Dynamic and parametric retrieval-augmented generation. In Proceedings of SIGIR ‘25, Padua, Italy. Association for Computing Machinery. [9] Tang, Xiaqiang, Li, Jian, Du, Nan, & Xie, Sihong. (2025). Adapting to non-stationary environments: Multiarmed bandit enhanced retrieval-augmented generation on knowledge graphs. In Proceedings of the AAAI Conference on Artificial Intelligence (Vol. 39, No. 12). AAAI-25 Technical Tracks 12. [10] Cheng, Sibo., Prentice, I. Colin, Huang., Yuhan, Huang., Jin, Yufang., Guo., Yi-Ke., Arcucci, Rossella. (2022). Data-driven surrogate model with latent data assimilation: Application to wildfire forecasting. Journal of Computational Physics, 464, 111302. [11] Charotia, Himanshi., Garg, Abhishek., Dhama, Gaurav., Maheshwari, Naman. (2021). Optimal latent space forecasting for large collections of short time series using temporal matrix factorization: A preprint. arXiv preprint arXiv:2112.08052v1. ] 15 Dec 2021. [12] Liu, Nelson F., Lin, Kevin, Hewitt, John, Paranjape, Ashwin, Bevilacqua, Michele, Petroni, Fabio, & Liang, Percy. (2023). Lost in the middle: How language models use long contexts. arXiv preprint arXiv:2307.03172. [13] Borgeaud, Sebastian., Mensch, Arthur., Hoffmann, Jordan., Cai, Trevor., Rutherford, Eliza., Millican, Katie., Van Den Driessche, George Bm., Lespiau, Jean-Baptiste., Damoc, Bogdan., Clark, Aidan. (2022). Improving language models by retrieving from trillions of tokens. In International Conference on Machine Learning (pp. 2206–2240). Proceedings of Machine Learning Research (PMLR). [14] Jiang, Zhengbao., Xu, Frank F., Gao, Luyu., Sun, Zhiqing., Liu, Qian., Dwivedi-Yu., Jane, Yang., Yiming, Callan, Jamie., Neubig, Graham. (2023). Active retrieval augmented generation. arXiv preprint arXiv:2305.06983. [15] Zhong, Zexuan., Lei, Tao., Chen, Danqi. (2022). Training language models with memory augmentation. arXiv preprint arXiv:2205.12674. [16] Lewis, Patrick., Perez, Ethan., Piktus, Aleksandra., Petroni, Fabio., Karpukhin, Vladimir., Goyal, Naman., Küttler, Heinrich., Lewis, Mike., Yih, Wen-tau., Rocktäschel, Tim., Riedel, Sebastian., Kiela, Douwe. (2020). Retrieval-augmented generation for knowledge-intensive NLP tasks. In: Proceedings of the 34th Conference on Neural Information Processing Systems (NeurIPS 2020), Vancouver, Canada. [17] Xu, Y., Cohen, S. B. (2018). Stock movement prediction from tweets and historical prices. In: Proceedings of the Association for Computational Linguistics (pp. 1970–1979). [18] Shi, L., Teng, Z., Wang, L., Zhang, Y., Binder, A. (2019). DeepClue: Visual interpretation of text-based deep stock prediction. IEEE Transactions on Knowledge and Data Engineering, 31(6), 1094–1108. [19] Cheng, D., Yang, F., Xiang, S., Liu, J. (2022). Financial time series forecasting with multi-modality graph neural network. Pattern Recognition, 121, 108218. [20] Xu, W., Liu, W., Wang, L., Xia, Y., Bian, J., Yin, J., Liu, T. (2021). HIST: A graph-based framework for stock trend forecasting via mining concept-oriented shared information. arXiv preprint arXiv:2110.13716.