[目的/意义] 人工智能（AI）正引发链式反应般的科学突破，引领新一轮科技革命和产业变革，图书文献情报领域如何利用AI技术提供智慧知识服务与智能情报系统是当前行业关注的焦点与热点。[方法/过程] 从图书情报行业内外综合分析AI技术与大数据为知识服务范式带来的新平台、新服务以及新机遇与新挑战，提出"AI技术+大数据"驱动的智慧知识服务生态体系建设的总体思路，从智慧数据、智慧中台与智慧服务3个层面共同构建"科情大脑"，提供覆盖科技管理、科技创新与社会学术信息环境的开放智慧知识服务生态环境。[结果/结论] 以中国科学院文献情报中心的文献情报数据湖、智能知识服务引擎、智慧知识发现、智慧知识管理、智能情报分析系统以及智能感知环境6个方面进行探索建设，取得有意义的成效。面向未来，阐明AI技术在面向大数据治理、细粒度知识识别、精准服务提供等方面，仍需要在数据、技术以及服务模式上进一步提升。
[Purpose/significance] Artificial Intelligence is triggering a chain reaction-like scientific breakthrough, leading a new round of scientific and technological revolution and industrial transformation, how to use AI to provide Intelligent Knowledge Services and Intelligent Information system is the current focus of attention and hot spot. [Method/process] This paper analyzed AI technology and big data from inside and outside the library and information domain to bring new platforms, new services and new opportunities and challenges to the knowledge service paradigm, and provided the idea of building the Intelligent Knowledge Service ecosystem driven by "AI technology and big data", and built the "science brain" method from the three levels of intelligent data, wisdom center and intelligence service, and provided an open intelligent knowledge service ecological environment covering science and technology management, scientific and technological innovation and social academic information environment. [Result/conclusion] As an exploration building of National Science Library Chinese Academy of Sciences, about Data Lake of Library and Information, Intelligent Knowledge Service engine, Intelligent Knowledge Discovery, Intelligent Knowledge Management, Intelligent Knowledge Analysis and Intelligent Sense Environment, and good results have been achieved. In the future, AI technology for big data government, fine-grained knowledge recognition, precision service and so on, still need to be further improved in data, technology and service models.
[1] 张冬荣, 钱力. "科技大数据与智能知识服务平台建设"专题序[J]. 数据分析与知识发现, 2019, 3(1):3-3.
[2] 柯平,邹金汇. 后知识服务时代的图书馆转型[J] 中国图书馆学报. 2019,45(1):4-17.
[3] 罗立群,李广建. 智慧情报服务与知识融合[J] 情报资料工作. 2019,40(2):87-94.
[4] 苏新宁. 新时代图书馆使命与未来图书馆学教育之思考[J].中国图书馆学报,2020(1):53-62.
[5] 张晓林. 颠覆性变革与后图书馆时代-推动知识服务的供给侧结构性改革[J]. 中国图书馆学报, 2018, 44(1):4-16.
[6] 钱力,谢靖,常志军,等. 基于科技大数据的智能知识服务体系研究设计[J]. 数据分析与知识发现, 2019, 3(1):4-14.
[7] 钱力, 张晓林, 王茜. 基于科技文献的研究设计指纹描述框架研究[J]. 大学图书馆学报, 2015, 33(1):14-20.
[8] 李广建, 江信昱. 论计算型情报分析[J]. 中国图书馆学报, 2018, 44(2):4-16.
[9] About us-polyplexus.com[EB/OL].[2020-10-06]. https://start.polyplexus.com/about-us/.
[10] wizdom.ai-intelligence for everyone[EB/OL].[2020-10-06]. https://www.wizdom.ai/#about.
[11] Digital science[EB/OL].[2020-10-06]. https://www.digital-science.com/.
[12] Research intelligence[EB/OL].[2020-10-06]. https://www.elsevier.com/research-intelligence.
[13] A free, AI-powered research tool for scientific literature[EB/OL].[2020-10-06]. https://www.semanticscholar.org/.
[14] Who's the most influential biomedical scientist? computer program guided by artificial intelligence says it knows|Science|AAAS[EB/OL].[2020-10-06]. http://www.sciencemag.org/news/2017/10/who-s-most-influential-biomedical-scientist-computer-program-guided-artificial.
[15] TSHITOYAN V, DAGDELEN J, WESTON L, et al. Unsupervised word embeddings capture latent knowledge from materials science literature[J]. Nature,2019, 571(7763), 95-98.
[16] KIM E, HUANG K, SAUNDERS A, et al. Materials synthesis insights from scientific literature via text extraction and machine learning[J]. Chemistry of materials,2017,29(21):9436-9444.
[17] ZHOU Q, TANG P, LIU S, et al. Learning atoms for materials discovery[J]. Proceedings of the National Academy of Sciences, 2018,115(28):E6411-E6417.
[18] KIYOHARA S, MIYATA T, TSUDA K, et al. Data-driven approach for the prediction and interpretation of core-electron loss spectroscopy[J]. Scientific reports, 20188(1):1-12.
[19] STOKES J M, YANG K, SWANSON K,et al. A deep learning approach to antibiotic discovery[J]. Cell, 2020,180(4):688-702.
[20] SEGLER M H, PREUSS M, WALLER M P. Planning chemical syntheses with deep neural networks and symbolic AI[J]. Nature, 2018,555(7698):604-610.
[21] GRANDA J M, DONINA L, DRAGONE V,et al. Controlling an organic synthesis robot with machine learning to search for new reactivity[J]. Nature,2018, 559(7714):377-381.
[22] POPOVA M, ISAYEV O, TROPSHA A. Deep reinforcement learning for de novo drug design[J]. Science advances, 2018,4(7):7885.
[23] RADOVIC A, WILLIAMS M, ROUSSEAU D, et al. Machine learning at the energy and intensity frontiers of particle physics[J]. Nature,2018, 560(7716):41-48.
[24] MATHURIYA A, BARD D, MENDYGRAL P, et al. CosmoFlow:using deep learning to learn the universe at scale[C]//SC18:international conference for high performance computing. Dallas Texas:IEEE, 2018:819-829.
[25] ZHANG Y G, GAJJAR V, FOSTER G, et al. Fast radio burst 121102 pulse detection and periodicity:a machine learning approach[J]. The astrophysical journal, 2018,866(2):149.
[26] BEHESHTIL A, BENATALLAN B, SHENG Q Z, et al. Intelligent knowledge lakes:the age of artificial intelligence and big data[C]//WISE 2020. Singpore:Springer, 2020.
[27] Connecting life science data to enable impactful analytics-Entellect|Elsevier[EB/OL].[2020-10-06]. https://www.elsevier.com/solutions/entellect.
[28] AI-assisted reagent selection and experiment design-BenchSci[EB/OL].[2020-10-06]. https://www.benchsci.com/.
[29] Your science assistant[EB/OL].[2020-10-06]. https://iris.ai/.
[30] 学术出版商Springer Nature出版了第一本由AI作家创作的书籍[EB/OL].[2020-10-06]. http://m.elecfans.com/article/906694.html.
[31] Transforming information into knowledge[EB/OL].[2020-10-06]. https://www.yewno.com/.
[32] Google's neural machine translation system:bridging the gap between human and machine translation[EB/OL].[2020-10-06]. https://arxiv.org/pdf/1609.08144.pdf.
[33] SciBERT:a pretrained language model for scientific text[EB/OL].[2020-10-06]. https://arxiv.org/abs/1903.10676.
[34] BioBERT:a pre-trained biomedical language representation model for biomedical text mining.[EB/OL].[2020-10-06]. https://arxiv.org/abs/1901.08746.
[35] Pre-training with Whole Word Masking for Chinese BERT[EB/OL].[2020-10-06]. https://arxiv.org/abs/1906.08101.
[36] 中文任务全面超越BERT:百度正式发布NLP预训练模型ERNIE[EB/OL].[2020-10-06]. https://zhuanlan.zhihu.com/p/59436589.
[37] 基于科技文献预训练模型[EB/OL].[2020-10-06]. http://sciengine.las.ac.cn/.
[38] 王颖,钱力,谢靖,等. 科技大数据知识图谱构建模型与方法研究[J].数据分析与知识发现,2019(1):15-26.
[39] 国家数字科技文献资源长期保存体系[EB/OL].[2020-10-06]. http://ndpp.ac.cn/.
[40] 科技大数据多元知识融合治理云平台.[EB/OL].[2021-02-16]. http://data.scholarin.cn.
[41] 科技大数据多元知识融合治理云平台[EB/OL].[2020-10-06].http://data.scholarin.cn/.
[42] 新型冠状病毒专题-知识服务与科研攻关交流平台[EB/OL].[2020-10-06]. https://ncov.scholarin.cn/.
[43] 大数据探索发现系统[EB/OL].[2020-10-06]. http://kgview.las.ac.cn/discover.
[44] 领域科技情报监测服务云平台[EB/OL].[2020-10-06].http://stmcloud.las.ac.cn/.
[45] 基于科技文献知识资源的AI引擎[EB/OL].[2020-10-06]. http://sciengine.las.ac.cn/.
[46] 深度语义挖掘工具[EB/OL].[2020-10-06]. http://finger.las.ac.cn/Page_sharingTool_tm_NER.html.
[47] 中国科学院文献情报中心知识服务门户[EB/OL].[2020-10-06]. http://www.las.ac.cn.
[48] 慧科研-智能随身科研助理[EB/OL].[2020-10-06]. http://scholarin.cn.
