赞
踩
1.模型发展
2.面临挑战
检索增强生成(Retrieval-Augmented Generation, RAG)是一种结合了信息检索和文本生成技术的自然语言处理(NLP)方法。这种方法利用大型语言模型(LLM)的生成能力,并结合了检索系统从大量数据中检索相关信息的能力。RAG的目标是生成既准确又具有信息量的文本,同时确保生成的文本与给定的查询或任务紧密相关。通过检索与输入查询最相关的信息片段,然后使用这些片段来指导、优化和丰富生成过程,RAG提高了生成文本的准确性和相关性。
RAG在自然语言处理领域中的应用广泛,涵盖了问答系统、机器翻译、内容创作、对话系统等多个方面,该技术的作用如下:
检索增强生成(RAG)相较于传统语言模型具有几个显著优点:
[1] Yunfan G, Yun X, Xinyu G, Kangxiang J, Jinliu P, Yuxi B, Yi D, Jiawei S, Haofen W, et al. Retrieval-Augmented Generation for Large Language Models: A Survey[J], CoRR, 2023, abs/2312.10997
[2] Deng C, Yan W, Lemao L, Shuming S, et al. Recent Advances in Retrieval-Augmented Text Generation[C], Annual International ACM SIGIR Conference on Research and Development in Information Retrieval, 2022: 3417–3419.
[3] Ruochen Z, Hailin C, Weishi W, Fangkai J, Do X L, Chengwei Q, Bosheng D, Xiaobao G, Minzhi L, Xingxuan L, Shafiq J, et al. Retrieving Multimodal Information for Augmented Generation: A Survey.[J], CoRR, 2023, abs/2303.10868: 4736-4756.
[4] Xin C, Di L, Xiuying C, Lemao L, Dongyan Z, Rui Y, et al. Lift Yourself Up: Retrieval-augmented Text Generation with Self Memory[J], CoRR, 2023, abs/2305.02437
[5] Zhihong S, Yeyun G, Yelong S, Minlie H, Nan D, Weizhu C, et al. Enhancing Retrieval-Augmented Large Language Models with Iterative Retrieval-Generation Synergy.[J], CoRR, 2023, abs/2305.15294: 9248-9274.
[6] Menglin X, Xuchao Z, Camille C, Guoqing Z, Saravan R, Victor R, et al. Hybrid Retrieval-Augmented Generation for Real-time Composition
Assistance[J], CoRR, 2023, abs/2308.04215
[7] Zachary L, Chenglu L, Wangda Z, Anoushka G, Owen H, Millie-Ellen P, Wanli X, et al. Retrieval-augmented Generation to Improve Math Question-Answering: Trade-offs Between Groundedness and Human Preference[J], CoRR, 2023, abs/2310.03184
Copyright © 2003-2013 www.wpsshop.cn 版权所有,并保留所有权利。