•  
  •  
 

Journal of Scientific Information Research

Keywords

large language model, academic innovation, chain-of-thought, text generation, systemic functional linguistics, automatic summarization

Abstract

[Purpose/significance] In response to the problems of non-standard expression in academic abstracts and lack of standardized methods for writing innovative highlights, this study explores an automatic generation method for innovative highlights based on the large language model. The aim is to provide researchers with diverse expression references and optimization suggestions, assist them in more accurately extracting and presenting research innovation points, improve the accuracy of academic innovation expression, and serve innovation content evaluation. [Method/process] Firstly, using the theory of ternary function in systemic functional linguistics, this paper analyzes the innovation highlights from multiple dimensions, such as innovation triggers, core terms, tense voice, sentence structure and abstract complementarity. Secondly, summarize the language expression characteristics of innovative highlights, and form a standardized language expression structure. Finally, the reasoning output of the large language model is optimized by designing the thinking chain according to the language characteristics to realize the automatic generation of innovative highlights. [Result/conclusion] Experimental results demonstrate that the chain-of-thought prompts designed in this study can effectively constrain the quality of text generated by large language models, with the generated research highlights performing well in terms of content richness, logical clarity, and expression accuracy. This research provides an automated solution to assist scholars in efficiently extracting key innovations and supports large-scale innovative highlight generation for publishing institutions.

First Page

90

Last Page

101

Submission Date

20-Oct-2025

Revision Date

14-Jan-2026

Acceptance Date

12-Feb-2026

Published Date

1-Oct-2026

Reference

[1] 周海晨，郑德俊，郦天宇. 学术全文本的学术创新贡献识别探索[J]. 情报学报，2020，39（8）：845-851.

[2] TADDIO A，PAIN T，FASSOS F F，et al. Quality of nonstructured and structured abstracts of original research articles in the British Medical Journal，the Canadian Medical Association Journal and the Journal of the American Medical Association[J]. CMAJ：Canadian Medical Association Journal，1994，150（10）：1611-1615.

[3] WANG X G，GUI H，LIU J Z. How authors evaluate the novelty of their articles：A comparative analysis of highlights in research articles[J]. Journal of Information Science，2023，51：1250-1269.

[4] 王一博，郭鑫，刘智锋，等. AI生成与学者撰写中文论文摘要的检测与差异性比较研究[J]. 情报杂志，2023，42（9）：127-134.

[5] RAUNAK V，SHARAF A，WANG Y，et al. Leveraging GPT-4 for Automatic Translation Post-Editing[J]. arXiv e-prints，arXiv：2305.14878.

[6] LIU S，HUANG X J. A Chinese question answering system based on GPT[C]//2019 IEEE 10th International Conference on Software Engineering and Service Science（ICSESS）. Beijing，2019：533-537.

[7] YANG W. Evaluative language and interactive discourse in journal article highlights[J]. English for Specific Purposes，2016，42：89-103.

[8] 索传军，于果鑫. 学术论文研究亮点的语言学特征与分布规律研究[J]. 图书情报工作，2020，64（9）：104-113.

[9] 杨思洛，程濛，莫莹莹. 学术论文研究亮点的语言特征与主题识别[J]. 图书馆论坛，2023，43（7）：26-37.

[10] 刘越，郑德俊，程为. 面向信息资源管理领域的学术创新特征分析[J]. 图书情报工作，2024，68（4）：83-96.

[11] WANG W M，SEE-TO E W K，LIN H T，et al. Comparison of Automatic Extraction of Research Highlights and Abstracts of Journal Articles[C]//Proceedings of the 2nd International Conference on Computer Science and Application Engineering. New York，2018：1-5.

[12] CAGLIERO L，LA QUATRA M. Extracting highlights of scientific articles：A supervised summarization approach[J]. Expert Systems with Applications，2020，160：113659.

[13] REHMAN T，SANYAL D K，CHATTOPADHYAY S，et al. Generation of Highlights from Research Papers Using Pointer-Generator Networks and Scibert Embeddings[J]. arXiv e-prints，arXiv：2302.07729.

[14] 张启睿，马青壮，左迅. 基于大语言模型的政策法规问答系统研究[J]. 网络安全与数据治理，2025，44（S1）：50-54.

[15] 鲁凯文，杨雅婷，董瑞，等. 基于局部偏好优化的大模型机器翻译方法研究[J/OL]. 计算机工程，1-13（2025-09-22）[2025-11-23]. https://doi.org/10.19678/j.issn.1000-3428.0252213.

[16] 王丽媛，孙瑾，陈青华，等. 基于DeepSeek和零样本提示的装备领域非结构化信息抽取方法研究[J/OL]. 指挥控制与仿真，1-8（2025-06-11）[2025-11-23]. https://link.cnki.net/urlid/32.1759.tj.20250610.2229.024.

[17] 朱丹浩，黄肖宇，李堯霖，等. 基于大语言模型的法律文本的自动摘要方法[J]. 数据分析与知识发现，2025，9（6）：35-46.

[18] 周晶. 基于大语言模型的文本辅助生成系统的设计与实现[D]. 青岛：山东大学，2024.

[19] 马君锡，王小刚. 大语言模型在社交媒体文本生成领域的应用研究：以小红书平台为例[J]. 新媒体研究，2024，10（6）：16-20.

[20] BROWN T B，MANN B，RYDER N，et al. Language Models are Few-Shot Learners[J]. arXiv e-prints，arXiv：2005.14165.

[21] WEI J，WANG X，SCHUURMANS D，et al. Chain-of-Thought Prompting Elicits Reasoning in Large Language Models[J]. arXiv e-prints，arXiv：2201.11903.

[22] 张烜. 基于大模型的专利权利要求书自动生成技术研究[D]. 哈尔滨：哈尔滨工业大学，2025.

[23] 赵景欣，王志强，武永亮，等. 基于实体信息增强思维链的政策长文本摘要方法[J/OL]. 计算机技术与发展，1-11（2025-10-23）[2025-11-23]. https://doi.org/10.20165/j.cnki.ISSN1673-629X.2025.0299.

[24] 李战子. 后疫情时代的功能语言学话语分析[J]. 外语研究，2020，37（5）：1-6，112.

[25] DEVLIN J，CAHNG M W，LEE K，et al. BERT：Pre-training of Deep Bidirectional Transformers for Language Understanding[J]. arXiv e-prints，arXiv：1810.04805.

[26] HALLIDAY M A K. Language as Social Semiotic：The Social Interpretation of Language and Meaning[M]. London：Edward Arnold，1978.

[27] M.A.K.韩礼德，姜望琪，付毓玲. 篇章、语篇、信息：系统功能语言学视角[J]. 北京大学学报（哲学社会科学版），2011，48（1）：137-146.

[28] 胡壮麟. 系统功能语言学的认知观[J]. 外语学刊，2014（3）：44-50.

[29] MEYER J G，URBANOWICZ R J，MARTIN P C N，et al. ChatGPT and large language models in academia：opportunities and challenges[J]. BioData Mining，2023，16：20.

[30] 吴娜，刘畅，刘江峰，等. AIGC驱动古籍自动摘要研究：从自然语言理解到生成[J]. 图书馆论坛，2024，44（9）：111-123.

[31] 崔金满，李冬梅，田萱，等. 提示学习研究综述[J]. 计算机工程与应用，2024，60（23）：1-27.

[32] KOJIMA T，GU S S，REID M，et al. Large Language Models are Zero-Shot Reasoners[J]. arXiv e-prints，arXiv：2205.11916.

[33] ZHU F，LIU Z，FENG F，et al. TAT-LLM：A Specialized Language Model for Discrete Reasoning over Financial Tabular and Textual Data[C]//Proceedings of the 5th ACM International Conference on AI in Finance. Brooklyn，2024：310-318.

[34] 薛嗣媛，任福继. 基于对话思维链指令微调的汉语写作智能评语生成研究[J/OL]. 计算机工程与应用，1-10（2025-03-12）[2025-11-23]. http://kns.cnki.net/kcms/detail/11.2127.TP.20250312.1144.006.html.

[35] BELTAGY I，LO K，COHAN A. SciBERT：A Pretrained Language Model for Scientific Text[C]//Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing. Hong Kong，2019：3615-3620.

[36] CHEN Y B，XU L H，LIU K，et al. Event Extraction via Dynamic Multi-Pooling Convolutional Neural Networks[C]//Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing（Volume 1：Long Papers）. Beijing，2015：167-176.

[37] MIHALCEA R，TARAU P. TextRank：Bringing order into texts[C]//Proceedings of the 2004 Conference on Empirical Methods in Natural Language Processing. Barcelona，2004：404-411.

[38] SALTON G，BUCKLEY C. Term-weighting approaches in automatic text retrieval[J]. Information Processing & Management，1988，24（5）：513-523.

[39] RUSH A M，CHOPRA S，WESTON J. A Neural Attention Model for Abstractive Sentence Summarization[J]. arXiv e-prints，arXiv：1509.00685.

[40] LEWIS M，LIU Y，GOYAL N，et al. BART：Denoising Sequence-to-Sequence Pre-training for Natural Language Generation，Translation，and Comprehension[C]//Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics. Online，2020：7871-7880.

[41] 其其日力格，斯琴图，王斯日古楞. 基于深度学习的自动文本摘要研究综述[J]. 计算机工程与应用，2025，61（18）：24-40.

[42] LIU A X，FENG B，XUE B，et al. Deepseek-V3 Technical Report[J]. arXiv e-prints，arXiv：2412.19437.

[43] YANG A，YANG B S，ZHANG B C，et al. Qwen2.5 Technical Report[J]. arXiv e-prints，arXiv：2412.15115.

[44] ACHIAM J，ADLER S，AGARWAL S，et al. Gpt-4 Technical Report[J]. arXiv e-prints，arXiv：2303.08774.

[45] 王俊超，樊可汗，霍智恒. 中文大模型生成式摘要能力评估[J]. 中文信息学报，2025，39（1）：1-15.

[46] FOSTIROPOULOS I，ITTI L. ABLATOR：Robust Horizontal-Scaling of Machine Learning Ablation Experiments[C]//Proceedings of the Second International Conference on Automated Machine Learning（PMLR）. Potsdam/Berlin，2023：19-1-15.

Digital Object Identifier (DOI)

10.19809/j.cnki.kjqbyj.2026.04.009

Download

Share

COinS