Exploration and Evolution of Technology Themes in AI-empowered Biomanufacturing Driven by Application Scenarios

doi:10.13560/j.cnki.biotech.bull.1985.2026-0227

Abstract

Abstract:

Objective Biological manufacturing is an important track for cultivating new quality productive forces, and the empowerment of AI technology has become a key factor in driving technological innovation in the field of biological manufacturing. Exploring the thematic relationships and evolutionary trends between AI- empowered technology and application scenarios in the field of biological manufacturing can help reveal the cross-fusion trend between technology and application scenarios, thereby accurately anchoring the future direction and potential of technological development. Method We proposed a knowledge-augmented BERTopic topic model framework for AI-enabled bio-manufacturing to elucidate the technological and application scenario themes, their associations, and evolution, thereby revealing the developmental trajectory of AI technologies tailored to application scenarios. Firstly, LLM prompt was used to recognize technical entities and scene entities in papers and patent texts, and the SBERT-Kmeans entity alignment method was employed to achieve semantic consistency in entity representation. Then, the BERTopic model was used to conduct topic analysis on the fused domain knowledge, identifying technology and application scenarios for research topics. Finally, we established the correlation between technology and application scenarios, introduced the time dimension to discovered the dynamic evolution trends of technology, application scenarios, and the “technology-scenario” theme, and analyzed the evolution trends of hot research topics in AI-enabled biomanufacturing. Result The field of AI-enabled biomanufacturing mainly includes 7 research topics on AI technology and 6 research topics on application scenarios. Machine learning models and neural network models occupy a core position, while new technologies such as AlphaFold model and Transformer model have received more attention in recent years. Drug discovery and design, as well as protein structure prediction and design, have emerged as future research hotspots and trends. Conclusion The knowledge-augmented BERTopic topic model helps to improve the accuracy and bias of domain theme recognition, and the research conclusions provide decision-making support for industrial planning and innovation strategy formulation.

Key words: biomanufacturing, AI technologies, application scenarios, BERTopic, large language models

LI Zhen-zhen, GAO Qian, ZHONG Yong-heng, WANG Hui, LIU Jia. Exploration and Evolution of Technology Themes in AI-empowered Biomanufacturing Driven by Application Scenarios[J]. Biotechnology Bulletin, doi: 10.13560/j.cnki.biotech.bull.1985.2026-0227.

Figures/Tables 7

Fig. 1 Technology route of this research

Fig. 2 Time series trends of AI-empowered scientific and technological paper publications and patent literature applications in the field of biomanufacturing

Fig. 3 Two‑dimensional visualization distribution map of the AI- empowered biomanufacturing fieldA: AI technologies; B: application scenarios

Fig. 4 Hierarchical clustering diagram of research themes in AI‑ empowered biomanufacturingA: AI technologies; B: application scenarios

Fig. 5 "Technology-Scene" theme correlation diagram in AI‑empowered biomanufacturing

Fig. 6 Evolutionary trends of research themes in AI-empowered biomanufacturing (From 2014 to 2025)A: AI technology topics; B: application scenario topics; C: “technology-scenario” topics

Fig. 7 Evolutionary trends of technology themes from the perspective of various application scenarios in AI‑empowered biomanufacturingA: Drug discovery and design; B: intelligent design and synthesis of biological components; C: intelligent control and optimization of biological processes; D: intelligent design and optimization of artificial cell factories; E: intelligent design and reconstruction of biological systems and biomolecules; F: biological text mining and knowledge discovery

References 33

[1]	张以恒. 中国哲学思想"道法术器"对生物制造的启示 [J]. 合成生物学, 2024, 5(6): 1231-1241.
	Zhang YH. The enlightenment of the Chinese philosophy “Tao-Fa-Shu-Qi” to industrial biomanufacturing [J]. Synth Biol J, 2024, 5(6): 1231-1241.
[2]	王钦宏, 马红武, 夏建业. AI驱动生物制造专刊序言 [J]. 生物工程学报, 2025, 41(3): I0001-I0008.
	Wang QH, Ma HW, Xia JY. Preface for special issue on AI-driven biomanufacturing [J]. Chin J Biotechnol, 2025, 41(3): I0001-I0008.
[3]	Lin Y, Wang XL, Yang J, et al. Core technology topic identification and evolution analysis based on patent text mining—a case study of unmanned ship [J]. Appl Sci, 2024, 14(11): 4661.
[4]	尹西明, 陈泰伦, 金珺. 引领未来产业发展的产业先导能力: 理论逻辑与过程机制 [J]. 科学学与科学技术管理, 2026, 47(1): 30-41.
	Yin XM, Chen TL, Jin J. Industrial anticipatory capabilities leading future industry development: theoretical logic and operational mechanism [J]. Sci Sci Manag S T Mon, 2026, 47(1): 30-41.
[5]	王益成, 蒋星宇, 秦晴, 等. 基于投融资事件的技术主题识别研究——以生物技术领域为例 [J]. 情报学报, 2025, 44(2): 234-245.
	Wang YC, Jiang XY, Qin Q, et al. Research on technology theme identification based on investment and financing events——using biotechnology as an example [J]. J China Soc Sci Tech Inf, 2025, 44(2): 234-245.
[6]	李佳恒, 王萍, 凌茗, 等. 基于政府资助企业产品开发项目的技术主题识别研究 [J]. 情报理论与实践, 2025, 48(10): 140-147.
	Li JH, Wang P, Ling M, et al. Research on technology topic identification based on government-funded enterprise product development projects [J]. Inf Stud Theory Appl, 2025, 48(10): 140-147.
[7]	李雪, 王婧朱, 刘心蕊, 等. 基于时序关联和价值评估的技术主题演化研究——以生物制药产业为例 [J]. 数字图书馆论坛, 2023, 19(6): 22-30.
	Li X, Wang JZ, Liu XR, et al. Evolution of technological topic based on temporal correlation and value evaluation: a case study of bio-pharmaceutical industry [J]. Digit Libr Forum, 2023, 19(6): 22-30.
[8]	刘春丽, 臧东宇, 陈爽. 科学—技术—产业关联测度与主题演化规律研究——以生物医药领域为例 [J]. 图书情报工作, 2024, 68(14): 95-116.
	Liu CL, Zang DY, Chen S. Research on the correlation measure and theme evolution law of science-technology-industry: taking the field of biomedicine as an example [J]. Libr Inf Serv, 2024, 68(14): 95-116.
[9]	Zheng XL, Wu Q, Luo XY, et al. Analyzing the evolutionary trajectory of technological themes based on the BERTopic model: a case study in the field of artificial intelligence [J]. PLoS One, 2025, 20(6): e0324933.
[10]	孙文晶, 马捷, 郝志远. 融合SBERT与自适应HDBSCAN算法的技术主题识别及演化研究——以人工智能领域为例 [J]. 情报理论与实践, 2026, 49(3): 139-149.
	Sun WJ, Ma J, Hao ZY. Research on technical topic identification and evolution by integrating SBERT and adaptive HDBSCAN algorithms: taking the field of artificial intelligence as an example [J]. Inf Stud Theory Appl, 2026, 49(3): 139-149.
[11]	刘湘贝, 严亚兰, 查先进. 基于BERTopic的人工智能应用场景主题建模研究 [J]. 图书与情报, 2025(3): 46-55.
	Liu XB, Yan YL, Zha XJ. Study on topic modeling of artificial intelligence application scenario based on BERTopic [J]. Libr Inf, 2025(3): 46-55.
[12]	叶选挺, 马诗敏, 王宇, 等. 场景驱动视角下我国智能医疗产业演化研究 [J]. 科技进步与对策, 2023, 40(24): 20-30.
	Ye XT, Ma SM, Wang Y, et al. The evolution of intelligent medical industry from the scenario-driven perspective [J]. Sci Technol Prog Policy, 2023, 40(24): 20-30.
[13]	谢俊杰, 孙希科, 王智琦, 等. 基于应用场景的未来技术识别 [J]. 情报杂志, 2024, 43(5): 97-105.
	Xie JJ, Sun XK, Wang ZQ, et al. Future technology identification based on application scenarios [J]. J Intell, 2024, 43(5): 97-105.
[14]	范惠明, 金莹. 元宇宙教育应用技术与工程教育深度学习: 技术主题识别与应用场景对接 [J]. 电化教育研究, 2025, 46(10): 55-62.
	Fan HM, Jin Y. Metaverse educational application technology and deep learning in engineering education: identification of technical themes and alignment with application scenarios [J]. E Educ Res, 2025, 46(10): 55-62.
[15]	苗红, 王浩桐, 李伟伟, 等. 面向应用场景的前沿技术识别方法 [J]. 情报杂志, 2025, 44(1): 95-103.
	Miao H, Wang HT, Li WW, et al. Frontier technology identification method for application scenarios [J]. J Intell, 2025, 44(1): 95-103.
[16]	Kusumaningrum R, Khoerunnisa SF, Khadijah K, et al. Exploring community awareness of mangrove ecosystem preservation through sentence-BERT and K-means clustering [J]. Information, 2024, 15(3): 165.
[17]	Vardakas G, Likas A. Global k-means++: an effective relaxation of the global k-means clustering algorithm [J]. Appl Intell, 2024, 54(19): 8876-8888.
[18]	Khodeir N, Elghannam F. Efficient topic identification for urgent MOOC Forum posts using BERTopic and traditional topic modeling techniques [J]. Educ Inf Technol, 2025, 30(5): 5501-5527.
[19]	He XH, Li JR, Shen SY, et al. AlphaFold3 versus experimental structures: assessment of the accuracy in ligand-bound G protein-coupled receptors [J]. Acta Pharmacol Sin, 2025, 46(4): 1111-1122.
[20]	王震宇, 朱学芳, 夏思洋, 等. 基于视觉语言Transformer的社交媒体命名实体识别研究 [J]. 现代情报, 2025, 45(9): 46-57.
	Wang ZY, Zhu XF, Xia SY, et al. Research on named entity recognition in social media based on vision-and-language transformers [J]. J Mod Inf, 2025, 45(9): 46-57.
[21]	Hua Y, Feng ZH, Song XN, et al. MMDG-DTI: Drug-target interaction prediction via multimodal feature fusion and domain generalization [J]. Pattern Recognit, 2025, 157: 110887.
[22]	Ullah A, Rehman B, Khan S, et al. An in silico multi-epitopes vaccine ensemble and characterization against nosocomial Proteus penneri [J]. Mol Biotechnol, 2024, 66(12): 3498-3513.
[23]	de Fátima Cobre A, Ara A, Alves AC, et al. Identifying 124 new anti-HIV drug candidates in a 37 billion-compound database: an integrated approach of machine learning (QSAR), molecular docking, and molecular dynamics simulation [J]. Chemom Intell Lab Syst, 2024, 250: 105145.
[24]	Abramson J, Adler J, Dunger J, et al. Addendum: Accurate structure prediction of biomolecular interactions with AlphaFold 3 [J]. Nature, 2024, 636(8042): E4.
[25]	Liu HB, Li DG, Wu H. lncLocator-imb: an imbalance-tolerant ensemble deep learning framework for predicting long non-coding RNA subcellular localization [J]. IEEE J Biomed Health Inform, 2024, 28(1): 538-547.
[26]	Dan A, Liu BC, Patil U, et al. Machine learning model-based design and model predictive control of a bioreactor for the improved production of mammalian cell-based bio-therapeutics [J]. Control Eng Pract, 2025, 156: 106198.
[27]	Yatipanthalawa BS, Wallace Fitzsimons SE, Horning T, et al. Development and validation of a hybrid model for prediction of viable cell density, titer and cumulative glucose consumption in a mammalian cell culture system [J]. Comput Chem Eng, 2024, 184: 108648.
[28]	Heidari F, Mahdavinejad M, Zolotovsky K, et al. Data-driven bio-integrated design method encoded by biocomputational real-time feedback loop and deep semi-supervised learning (DSSL) [J]. J Build Eng, 2024, 98: 110923.
[29]	齐显尼, 何泓兴, 孔泽曦, 等. 微生物细胞工厂驱动新质蛋白生物制造 [J]. 中国食品学报, 2025, 25(9): 16-30.
	Qi XN, He HX, Kong ZX, et al. Neoprotein biomanufacturing via designer microbial cell factories [J]. J Chin Inst Food Sci Technol, 2025, 25(9): 16-30.
[30]	Zeng JY, Song KF, Wang JQ, et al. Characterization and optimization of 5' untranslated region containing poly-adenine tracts in Kluyveromyces marxianus using machine-learning model [J]. Microb Cell Fact, 2024, 23: 7.
[31]	Ahmed NY, Alsanousi WA, Hamid EM, et al. An efficient deep learning approach for DNA-binding proteins classification from primary sequences [J]. Int J Comput Intell Syst, 2024, 17: 88.
[32]	Hou G, Jian YH, Zhao QQ, et al. Language model based on deep learning network for biomedical named entity recognition [J]. Methods, 2024, 226: 71-77.
[33]	Wang YQ, Zuo JM, Duan C, et al. Large language models assisted multi-effect variants mining on cerebral cavernous malformation familial whole genome sequencing [J]. Comput Struct Biotechnol J, 2024, 23: 843-858.