黄水清,许乾坤,王东波,王林旭

• 1:南京农业大学信息管理学院
• 2:南京农业大学人文与社会计算研究中心

摘要(Abstract)：

[目的/意义]人工智能驱动的科学研究(artificial intelligence for science，AI4S)正推动科研从传统假设驱动向数据、模型协同驱动的新范式演进。系统考察AI4S赋能信息资源管理的演进脉络、方法体系、实践场景与治理挑战在当前很有必要。[方法/过程]AI4S赋能信息资源管理在方法与实践层面均已取得长足的进步，大语言模型在检索增强生成、多智能体协同、多模态语义对齐、知识图谱融合、文献情报获取、同行评审、学术出版与跨语言科学交流等多种场景中重塑科研生产力。然而，模型“幻觉”与知识库污染、算法操控与系统性偏见、AIGC著作权争议、数据垄断与信息茧房等问题，使学术伦理与科研治理面临新的张力。[结果/结论]基于此，信息资源管理学科应主动从“文献服务的支持学科”转向“科研创新的基础设施学科”，通过构建开放可信的科学数据底座、完善人机协同的学术规范、强化算法监管与内容溯源机制，承担起AI时代高质量数据供给者、知识生命周期治理者与学术伦理守护者的新使命。

关键词(KeyWords)： AI for Science;科研范式;信息资源管理;大语言模型

Abstract:

Keywords:

基金项目(Foundation): 国家社会科学基金重大项目“先秦诸子典籍知识库建设及词典编纂”（项目编号：22&ZD262）研究成果之一

作者(Author): 黄水清,许乾坤,王东波,王林旭

参考文献(References)：

• [1]Hey T,Tansley S,Tolle K.Jim Gray on eScience:a transformed scientific method[M]//Hey T,Tansley S,Tolle K.The fourth paradigm:data-intensive scientific discovery.Washington:Microsoft Research,2009:xvii-xxxi.
• [2]Wang H C,Fu T F,Du Y Q,et al.Scientific discovery in the age of artificial intelligence[J].Nature,2023,620(7972):47-60.
• [3]Karpatne A,Deshwal A,Jia X W,et al.AI-enabled scientific revolution in the age of generative AI:Second NSF workshop report[J].npj Artificial Intelligence,2025,1(1):article No.18.
• [4]Jumper J,Evans R,Pritzel A,et al.Highly accurate protein structure prediction with AlphaFold[J].Nature,2021,596(7873):583-589.
• [5]Biesenbender S,Petersohn S,Thiedig C.Using current research information systems (CRIS) to showcase national and institutional research (potential):Research information systems in the context of Open Science[J].Procedia Computer Science,2019,146:142-155.
• [6]Materska K.Research information management in the context of open science[J].Zagadnienia Informacji Naukowej-Studia Informacyjne,2019,57(1A):37-45.
• [7]李伦,刘梦迪.人工智能驱动的科学研究范式革命:态势与未来[J].探索与争鸣,2024(10):143-151,180.
• [8]Dudley C,Magdaleno R,Harding C,et al.Simulation as supervision:Mechanistic pretraining for scientific discovery[PP/OL].V4.arXiv (2026-04-14)[2026-06-19].https://arxiv.org/abs/2507.08977.
• [9]Qian E,Kang D,Sella V,et al.Multifidelity linear regression for scientific machine learning from scarce data[PP/OL].V2.arXiv (2024-07-01)[2026-06-19].https://arxiv.org/abs/2403.08627.
• [10]Muckley E S,Saal J E,Meredig B,et al.Interpretable models for extrapolation in scientific machine learning[J].Digital Discovery,2023,2(5):1425-1435.
• [11]Karniadakis G E,Kevrekidis I G,Lu L,et al.Physicsinformed machine learning[J].Nature Reviews Physics,2021,3(6):422-440.
• [12]Hu Z Y,Kawaguchi K,Zhang Z Q,et al.Tackling the curse of dimensionality in fractional and tempered fractional PDEs with physics-informed neural networks[J].Computer Methods in Applied Mechanics and Engineering,2024(432):117448.
• [13]Smyl D.Implicit and explicit treatments of model error in numerical simulation[J].Archives of Computational Methods in Engineering,2026,33(3):4501-4512.
• [14]Cybenko G.Approximation by superpositions of a sigmoidal function[J].Mathematics of Control,Signals and Systems,1989,2(4):303-314.
• [15]Han J Q,Cen J C,Wu L M,et al.A survey of geometric graph neural networks:Data structures,models and applications[J].Frontiers of Computer Science,2025,19(11):1911375.
• [16]Kramer S,Cerrato M,Brugger J,et al.Automated scientific discovery:From equation discovery to autonomous discovery systems[J].Machine Learning,2026,115(5):article No.109.
• [17]Shavit Y,Agarwal S,Brundage M,et al.Practices for governing agentic AI systems[R/OL].(2023-12-14)[2026-07-23].https://cdn.openai.com/papers/practicesfor-governing-agentic-ai-systems.pdf.
• [18]Ahart J.Will AI ever win its own Nobel?Some predict a prize-worthy science discovery soon[J].Nature,2025,646(8084):276-278.
• [19]李自拓,孙建彬,陈广州,等.大语言模型幻觉检测方法综述[J].计算机研究与发展,2026,63(1):123-146.
• [20]李子骏,肖辉,李雪峰.面向知识密集型任务的检索增强生成技术综述[J].微电子学与计算机,2025,42(10):48-65.
• [21]赵鑫,窦志成,文继荣.大语言模型时代下的信息检索研究发展趋势[J].中国科学基金,2023,37(5):786-792.
• [22]文森,钱力,胡懋地,等.基于大语言模型的问答技术研究进展综述[J].数据分析与知识发现,2024,8(6):16-29.
• [23]Wu W L,Wang H F,Li B H,et al.MultiRAG:Aknowledge-guided framework for mitigating hallucination in multi-source retrieval augmented generation[C]//Proceedings of the 2025 IEEE 41st International Conference on Data Engineering (ICDE).Piscataway:IEEE,2025:3070-3083.
• [24]李鑫,于汉超.人工智能驱动的生命科学研究新范式[J].中国科学院院刊,2024,39(1):50-58.
• [25]夏冰,胡梦良,袁天,等.多智能体协作工作流综述[J].软件导刊,2026,25(3):1-8.
• [26]董益宏,姜雪,钱家如,等.基于大语言模型智能体的代码生成综述[J/OL].软件学报,2026:1-27.(2026-04-02)[2026-07-22].https://doi.org/10.13328/j.cnki.jos.007593.
• [27]Kim Y,Gu K,Park C,et al.Towards a science of scaling agent systems[PP/OL].V3.arXiv (2026-04-08)[2026-07-23].https://arxiv.org/abs/2512.08296.
• [28]Cemri M,Pan M Z,Yang S Y,et al.Why do multiagent LLM systems fail?[C]//Proceedings of the 39th Conference on Neural Information Processing Systems.Red Hook:Curran Associates,2025:1-47.
• [29]焦明海,樊本航,王静,等.基于强化策略反馈的多模态自适应实体识别方法[J].计算机研究与发展,2026,63(2):294-304.
• [30]Wang X Q,Cui Y M,Yao X,et al.ChartHal:A finegrained framework evaluating hallucination of large vision language models in chart understanding[PP/OL].V1.arXiv (2025-09-22)[2026-07-23].https://arxiv.org/abs/2509.17481.